surface engineering of zrcual-based bulk metallic glasses ......e-mail: [email protected] technical...

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: May 11, 2021

Surface Engineering of ZrCuAl-based Bulk Metallic Glasses by Gaseous Oxidizing

Haratian, Saber

Publication date:2020

Document VersionPublisher's PDF, also known as Version of record

Link back to DTU Orbit

Citation (APA):Haratian, S. (2020). Surface Engineering of ZrCuAl-based Bulk Metallic Glasses by Gaseous Oxidizing.Technical University of Denmark.

https://orbit.dtu.dk/en/publications/24bd1776-1478-4a99-be1c-2589c7d268e3

PhD Thesis

DTU Mechanical EngineeringDepartment of Mechanical Engineering

Surface Engineering of ZrCuAl-based Bulk Metallic Glasses by Gaseous OxidizingSaber Haratian

SURFACE ENGINEERING OF

ZRCUAL-BASED BULK

METALLIC GLASSES BY

GASEOUS OXIDIZING

A THESIS SUBMITTED TO THE TECHNICAL UNIVERSITY OF

DENMARK FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN THE

DEPARTMENT OF MECHANICAL ENGINEERING

May 2020

By

Saber Haratian

Technical University of Denmark

Department of Mechanical Engineering

Section of Materials and Surface Engineering

i

Surface Engineering of ZrCuAl-based Bulk Metallic Glasses by Gaseous

Oxidizing

A doctoral thesis by

Saber Haratian

E-Mail: [email protected]


Department of Mechanical Engineering - Section of Materials and Surface Engineering

Building 425, room 111

DK-2800 Kgs. Lyngby

Principal Supervisor:

Prof. Marcel A.J. Somers




Produktionstorvet


DK-2800 Kgs. Lyngby

Co-Supervisors:

Associate Prof. Thomas L. Christiansen




Produktionstorvet


DK-2800 Kgs. Lyngby

Senior researcher Matteo Villa




Produktionstorvet


DK-2800 Kgs. Lyngby

Copyright: Reproduction of this publication in whole or in part must include the

customary bibliographic citation, including author attribution,

Surface engineering of ZrCuAl-based bulk metallic glasses by

gaseous oxidizing.

Published by: Department of Mechanical Engineering , Section of Materials and

Surface Engineering , Produktionstorvet, Building 425, DK-2800

Kgs. Lyngby

mailto:[email protected]

ii

To my beloved parents, Zohreh and Hossein,

my adored sister, Sara and her lovely daughter, Saba

iii

Preface

This doctoral thesis is submitted in partial fulfilment of the requirements for achieving the

degree of PhD at the Department of Mechanical Engineering at the Technical University

of Denmark (DTU). The project was funded by Villum Foundation under grant number

13253 and carried out at the Department of Mechanical Engineering, Section of Materials

and Surface Engineering, during the period from March 1st 2017 to May 31th 2020. The

project was supervised by Professor Marcel A.J. Somers, Associate Professor Thomas L.

Christiansen and Senior researcher M. Villa from the Department of Mechanical

Engineering, Section of Materials and Surface Engineering, at the Technical University of

Denmark (DTU).

Saber Haratian

Kongens Lyngby, Denmark, May 31th, 2020

iv

Abstract

Bulk metallic glasses (BMGs) are disordered metastable solids, wherein the composing

atoms are distributed randomly and in contrast to crystalline metals do not exhibit

translational periodic arrangement. The lack of crystalline defects in their atomic structure

makes them a potent metallic material, exhibiting exceptional mechanical properties, such

as high yield strength and high elastic limit. On the other hand, most of the BMGs fail

catastrophically when subjected to an external force, thus demonstrating poor ductility (and

plasticity). To overcome their intrinsic brittleness, a novel approach of surface-

compositional modification treatment by introducing oxygen atoms in their surface region

is hypothesized. This is considered to be an effective strategy to introduce compressive

residual stresses in the surface of BMGs as well as making their surface more crack-

resistant, which can potentially decelerate the surface crack-initiation. To this end, the

present doctoral thesis investigates the feasibility of surface engineering of multicomponent

ZrCuAl-based BMGs employing gaseous oxidizing. At the same time, a mechanistic study

is carried out to clarify the thermal oxidation behavior of this class of advanced materials.

Several aspects concerning the oxidation of three multicomponent ZrCuAl-based

BMGs, i.e. Zr48Cu36Al8Ag8, (Zr55Cu30Al10Ni5)98Er2, and Zr51.3Cu31.3Al8.5Ni4Ti4.9 BMGs

within this Ph.D. work are studied, including:

- The surface response of the highly-oxidizable ZrCuAl-based BMGs exposed to the

different controlled oxidizing environments at the temperature below their glass

transition.

- The surface microstructural evolution of these BMGs during air-oxidation.

- The effect of oxygen dissolution on the surface hardness of these materials.

- The determination of the state of stress (and strain) developed in the BMGs’ surface

region as a result of the ingression of oxygen.

- The possible stress relaxation mechanisms in oxidized BMGs.

In order to address these points, the thermal oxidation is followed in-situ and ex-

situ using X-ray diffraction (XRD) and thermogravimetry techniques, respectively. The

deliberate surface microstructural changes induced by oxygen dissolution at various

oxidizing conditions is investigated using a combination of different (post) microscopical

characterization techniques. Moreover, to provide a better understanding of the effect of

oxidation on the build-up of (compressive) residual stresses in the near-surface zone of the

BMGs, two independent techniques are applied, i.e. in-situ (and ex-situ) XRD lattice strain

(sin2ψ) analysis and an incremental ring-core focused ion beam milling combined with a

digital image correlation algorithm (FIB-DIC).

The main results of this Ph.D. project are categorized into four types of experimental

studies, each discussing the oxidation-induced microstructural changes and the

corresponding oxidation mechanisms comprehensively. The first study aims at exploring

the in-situ oxide phase evolution in the surface region of Zr48Cu36Al8Ag8 BMG exposed to

v

air at atmospheric conditions. The results of this work lead to the discovery of a self-repair

mechanism of microcracks (and shear band decoration) during oxidation of the investigated

BMG. The occurrence of this surprising effect emerges as crystallization and segregation

of the noble elements (Cu and Ag) within the free surfaces developed by oxidation. This

phenomenon is explained in terms of the development of compressive residual stresses as

a consequence of volume expansion associated with nano-crystalline ZrO2 formation. The

second study addresses the surface hardening of (Zr55Cu30Al10Ni5)98Er2 BMG in controlled

oxidizing atmospheres, imposing an extremely low and an extremely high oxygen partial

pressure. The results demonstrate that the surface hardness of the BMG can effectively be

improved by incorporating oxygen at elevated temperatures. Following this, the third study

elucidates the correlation between the observed oxidation-induced microstructural features

and the residual stresses developed in the surface of the (Zr55Cu30Al10Ni5)98Er2 BMG during

air-oxidation using in-situ XRD technique. In addition, stress relaxation mechanisms are

discussed in relation to crack formation perpendicular to the surface plane, caused by the

development of tensile stresses in the BMG underneath the oxidation zone and Cu (and Ni)

outward diffusion. Eventually, the fourth study focuses on the determination of

(compressive) residual stresses (and strains) developed in the surface of the

Zr51.3Cu31.3Al8.5Ni4Ti4.9 oxidized at an extremely low oxygen potential using ex-situ XRD

and FIB-DIC methods. The measurements yield comparable results and reveal the presence

of compressive residual stresses of about ~1.4-1.5 GPa, corresponding to approximately

macro-strains of ~0.45-0.50% in a surface-engineered BMG. Stress relaxation is

accomplished by shear band formation in the underlying BMG.

vi

Resumé

Bulk metallisk glas (BMG) er metastabile uordnede fast-fase materialer, hvor alle atomerne

er fordelt tilfældigt, uden nogen translatorisk periodisk ordning som ellers kendetegner

krystallinske metaller. Manglen på krystallinske defekter i deres atomstruktur gør dem til

”potente” metalliske materialer der udviser ekstraordinære mekaniske egenskaber såsom

høj flydespænding og høj elasticitetsgrænse. Dog lider de fleste BMG af pludselig brud,

når de udsættes for belastning, hvilket skyldes deres lave/dårlige duktilitet (og plasticitet).

For at imødekomme at BMG er skøre materialer undersøges her en ny metode til

overflademodificering som består i at introducere oxygenatomer i overfladen. Dette anses

som en effektiv strategi til at introducere trykspændinger i overfladen af BMG materialerne

samt gøre deres overflade mere revnebestandighed, hvilket potentielt kan bremse

revneåbning i overfladen. Til den ende undersøges i herværende Ph.D. afhandling

mulighederne for at benytte overfladebehandling af multikomponent ZrCuAl-baserede

BMG ved brug af gas-baseret oxidation. Samtidig gennemføres en mekanistisk

undersøgelse for at afklare hvordan denne klasse af avancerede materialer opfører sig ved

termisk oxidation.

Flere aspekter vedrørende oxidation af tre multikomponent ZrCuAl-baserede BMG

materialer, nemlig Zr48Cu36Al8Ag8, (Zr55Cu30Al10Ni5)98Er2, og Zr51.3Cu31.3Al8.5Ni4Ti4.9,

undersøges i denne Ph.D.:

- Overfladeresponsen af de stærkt reaktive (”oxiderbare”) ZrCuAl-baserede BMG

materialer, når de udsættes for forskellige kontrollerede oxidationsmiljøer ved

temperaturer under glasovergangstemperaturen.

- Udvikling af mikrostrukturen i overfladen af disse BMG materialer ved oxidation i

luft.

- Effekten af iltopløsning på overfladehårdheden af disse materialer.

- Bestemmelse af spændinger (og tøjning) som opstår i overfladen af BMG

materialerne som et resultat af indtrængning/indsætning af ilt.

- De mulige mekanismer for spændingsrelaksation i oxiderede BMG materialer.

For at adressere disse punkter følges den termiske oxidation in-situ og ex-situ under

anvendelse af henholdsvis røntgendiffraktion (XRD) og termogravimetri-teknikker. De

bevidste ændringer i overflade-mikrostrukturen induceret ved iltopløsning ved forskellige

oxidationsbetingelser undersøges ved anvendelse af en kombination af forskellige (post)

mikroskopiske karakteriseringsteknikker. Derudover, anvendes to uafhængige teknikker,

nemlig in-situ (og ex-situ) XRD analyse af tøjning (sin2ψ) og en trinvis ring-core fokuseret

ionstråle ”fræsning” kombineret med digital billedkorrelationsalgoritme (FIB-DIC) for at

opnå en bedre forståelse af virkningen af oxidation på opbygningen af (kompressive)

restspændinger i nær-overfladen af BMG materialerne.

vii

De vigtigste resultater fra denne Ph. D. afhandling kategoriseres i fire typer af

eksperimentelle undersøgelser som hver - i dybden - diskuterer de mikrostrukturelle

ændringer og tilhørende mekanismer som oxidation medfører. Det første studie undersøger

fase-udvikling in-situ i overfladen af Zr48Cu36Al8Ag8 BMG udsat for luftoxidation ved

atmosfæriske betingelser. Resultater af dette arbejde har ført til opdagelsen af en selv-

reparerende mekanisme af opståede mikro-revner (og shear bands) i BMG materialer ved

oxidation. Denne overraskende effekt fremkommer i form af krystallisation og segregering

af ædle elementer (Cu og Ag) i de frie overflader som dannes ved oxidation. Dette fænomen

kan forklares ud fra fremkomsten af trykspændinger som opstår som konsekvens af den

volumenudvidelse som er forbundet med dannelsen af nano-krystallinsk ZrO2. Det andet

studie adresserer overfladehærdning af (Zr55Cu30Al10Ni5)98Er2 BMG i kontrollerede

oxidations-atmosfærer som påtrykker enten et meget lavt eller meget højt ilt-partialtryk.

Resultaterne demonstrerer, at overfladehårdheden af BMG kan øges markant ved at

inkorporere ilt ved forøgede temperaturer. I forlængelse af dette, kaster det tredje studie lys

på sammenhængen mellem de oxidations-inducerede mikrostrukturelle features og

restspændingerne som opstår i overfladen af (Zr55Cu30Al10Ni5)98Er2 BMG under

luftoxidation ved brug af in-situ XRD teknik. Derudover bliver spændings-

relaksationsmekanismer diskuteret i forbindelse med revnedannelser som opstår vinkelret

på overfladen som konsekvens af fremkomsten af trækspændinger i BMG materialet under

oxidationszonen og pga. diffusion af Cu (og Ni) mod overfladen. Slutteligt fokuserer det

fjerde studie på bestemmelse af (tryk)spændinger (og tøjninger), som opstår i overfladen

på Zr51.3Cu31.3Al8.5Ni4Ti4.9 under oxidation ved ekstremt lavt ilt-partialtryk, ved hjælp af

ex-situ XRD og FIB-DIC metoder. Målingerne giver sammenlignelige resultater og viser

tilstedeværelsen af trykspændinger i størrelsesordenen ~ 1.4-1.5 GPa som omtrentlig svarer

til makro-tøjninger på ~ 0.45-0.50% i et overfladebehandlet BMG. Spændingsrelaksation

opnås ved shear band dannelse i det underliggende BMG.

viii

Acknowledgements

Foremost, I would like to express my appreciation and sincere gratitude to my principal

supervisor Prof. Marcel A.J. Somers, for his well-informed guidance, valuable pieces of

advice, and boundless support throughout this work. I am grateful to him for giving me the

opportunity to grow in his research group and broaden my scientific horizons. His scientific

spirit, insight, and intuition which always inspired me were invaluable. I could not have

imagined a better supervisor and mentor for my Ph.D. studies. Marcel, thank you so much

indeed for your patience and respectful attention during these years. It was a privilege to

carry out my Ph.D. studies on this exciting and challenging project under your supervision.

I also would like to express my sincere appreciation to my co-supervisors Associate

Prof. Thomas L. Christiansen and Senior researcher Matteo Villa for their support, valuable

and constructive suggestions during my Ph.D. studies. Their willingness to give their time

so generously is greatly appreciated. Thomas and Matteo, thank you for sharing your

experiences and always being there to show me the right direction. Thank you so much for

providing me your bright ideas and encouraging me to explore more and dig deeper into

the subjects.

I am very grateful to Flemming B. Grumsen for his technical knowledge and

continuous support in advanced materials characterization and constructive discussion

during our meetings, whenever they were required.

I extend my sincere thanks to Dr. Frank Niessen who has not only been a great

colleague but also a nice and supportive friend especially during my research visit at the

electron microscopy (EMC) centre of University of Wollongong (UOW). A part of the

Ph.D. project on the determination of surface strain relaxation (and residual stresses) in

oxidized BMG using incremental ring-core FIB-DIC technique would not have been

successful without his intelligence, knowledge, and enthusiasm. He is highly

acknowledged for supporting me in the process of developing the FIB-DIC technique and

optimizing the method parameters in the case of this project. Besides, I would like to

acknowledge Prof. Elena Pereloma for hosting me as a visiting research fellow at the UOW

electron microscopy centre. In addition, I wish to thank Dr. Mitchell J. B. Nancarrow

(UOW-EMC) for discussion, technical support, and experimental suggestions during my

external research stay.

I am very thankful to the Villum Fonden for the research-funding under grant

number 13253, and I wish to thank our project partners at Aalborg University including

Prof. Morten M. Smedskjær, Prof. Yuanzheng Yue, Dr. Kacper Januchta, and Dr. Malwina

Stepniewska for valuable discussion during the project meetings.

I further express my thanks to all my colleagues and friends at the Section of

Materials and Surface Engineering for their support and creating a positive working

environment. In particular, I wish to thank Gitte Salomon for administrative support,

ix

Steffen S. Munch, and Niklas Brinckman Gammeltoft-Hansen for support in the

metallographic laboratory. My special thanks go to some wonderful people at shared office

111 in building 425 who are still here or left for new adventures, including, Abhijeet Yadav,

Chiara Tibollo, Emilie H. Valente, Dr. Felix Lampert, Dr. Frank Niessen, Jacob O. Nielsen,

Konstantin V. Werner, Dr. Nicolai Y. Juul, and Dr. Sebastian N. B. Villadsen for making

the office such a nice working place.

Lastly, I would like to express my sincere indebtedness and gratitude to my dearest

friends for their continuous love throughout my life.

Last but not least, I would like to express my sincere thanks to my beloved family

for their endless support and love throughout my life and academic career. They have

always been and will be the reason for me to stand up and overcome all the challenges I

meet.

x

List of articles and conference contributions

Peer-reviewed articles included in this thesis:

1. S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, M. A.J. Somers, Self-

repair by stress-induced diffusion of the noble elements during oxidation of

Zr48Cu36Al8Ag8 bulk metallic glass, Scri. Mater. 164 (2019) 126–129. (Published)

2. S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, M.A.J. Somers, Surface

hardening by gaseous oxidizing of (Zr55Cu30Al10Ni5)98Er2 bulk metallic glass, J.

Alloys Comp., 800 (2019) 456–461. (Published)

3. S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, and M. A.J. Somers,

Stress, stress relaxation and self-healing in (Zr55Cu30Al10Ni5)98Er2 bulk metallic

glass during air-oxidation. (Ready for submission)

4. S. Haratian, F. Niessen, F. B. Grumsen, M. J. B. Nancarrow, E. Pereloma, M. Villa,

T. L. Christiansen, M. A.J. Somers, Strain, stress, and stress relaxation in oxidized

ZrCuAl-based bulk metallic glass. (Ready for submission)

Conference contributions in chronological order:

1. Oral presentation, 25th International Symposium of Metastable, Amorphous and

Nanostructured Materials, ISMANAM 2018, Rome, Italy, July 2nd

2. Oral presentation, 15th International Conference on the Physics of Non-Crystalline

Solids & 14th European Society of Glass Conference, PNCS-ESG 2018, Saint Malo,

July 9th

3. Oral presentation, 26th International Symposium of Metastable, Amorphous and

Nanostructured Materials, ISMANAM 2019, Chennai, India, July 8th

xi

Table of contents

Preface................................................................................................................................ iii

Abstract ............................................................................................................................... iv

Resumé ................................................................................................................................ vi

Acknowledgements .......................................................................................................... viii

List of articles and conference contributions ....................................................................... x

Table of contents ................................................................................................................. xi

List of abbreviations and symbols ..................................................................................... xv

1 Introduction .................................................................................................................. 1

1.1 Research framework and hypothesis ..................................................................... 1

1.2 Motivation and research objectives ....................................................................... 3

1.3 Scientific and experimental approach ................................................................... 4

1.4 Outline ................................................................................................................... 5

2 Background ................................................................................................................... 7

2.1 Metallic Glasses (MGs) ......................................................................................... 7

The concept of glass formation ...................................................................... 8

Development of bulk metallic glasses (BMGs) ........................................... 12

Fabrication methods ..................................................................................... 14

Functional properties of (bulk) metallic glasses .......................................... 16

Application fields of (bulk) metallic glasses ............................................... 18

2.2 Ductility and toughness enhancement of BMGs ................................................. 20

2.3 Oxidation behavior of metallic glasses ............................................................... 24

3 Materials and experimental methods .......................................................................... 37

3.1 Materials and gaseous oxidizing treatment ......................................................... 37

xii

3.2 Characterization .................................................................................................. 39

X-ray diffraction .......................................................................................... 39

Microscopical characterization .................................................................... 43

3.3 Residual stress determination .............................................................................. 51

Ex-situ (and in-situ) X-ray sin2ψ method .................................................... 52

Incremental ring-core focused ion beam (FIB) milling and digital image

correlation (DIC) ........................................................................................................ 55

4 Summary of results ..................................................................................................... 65

4.1 Manuscript I ........................................................................................................ 65

4.2 Manuscript II ....................................................................................................... 66

4.3 Manuscript III ...................................................................................................... 67

4.4 Manuscript IV ..................................................................................................... 68

5 Manuscript I ................................................................................................................ 71

5.1 Introduction ......................................................................................................... 72

5.2 Materials and methods ........................................................................................ 73

5.3 Results and interpretation .................................................................................... 74

In-situ X-ray diffraction ............................................................................... 74

Electron microscopical characterization of the oxide scale ......................... 75

5.4 Discussion ........................................................................................................... 79

5.5 Conclusion ........................................................................................................... 81

6 Manuscript II .............................................................................................................. 87

6.1 Introduction ......................................................................................................... 88

6.2 Materials and methods ........................................................................................ 89

6.3 Results and interpretation .................................................................................... 90

X-ray diffraction of as-cast and thermochemically oxidized BMG ............. 90

xiii

Cross-sectional hardness measurement........................................................ 91

Microstructure characterization of the oxidized zones ................................ 92

6.4 Discussion ........................................................................................................... 94

6.5 Conclusion ........................................................................................................... 97

7 Manuscript III ........................................................................................................... 103

7.1 Introduction ....................................................................................................... 104

7.2 Materials and Methods ...................................................................................... 105

Sample preparation and oxidation experiments ......................................... 105

Microscopical characterization of the oxidation zones .............................. 107

7.3 Results and interpretation .................................................................................. 107

Characterization of as-cast and as-oxidized BMG samples ....................... 107

In-situ XRD air-oxidation .......................................................................... 108

Microscopical characterization of the oxidation zones .............................. 109

Evaluation of the lattice strain and residual stress profiles ........................ 113

7.4 Discussion ......................................................................................................... 118

7.5 Conclusion ......................................................................................................... 120

8 Manuscript IV ........................................................................................................... 125

8.1 Introduction ....................................................................................................... 126

8.2 Experimental details and procedures................................................................. 128

Sample preparation and characterization ................................................... 128

XRD lattice strain determination ............................................................... 128

FIB-DIC measurement ............................................................................... 129

8.3 Results and interpretation .................................................................................. 131

X-ray diffraction and microscopical characterization ................................ 131

Evaluation of X-ray lattice strains with the sin2ψ method ......................... 133

xiv

Surface strain relaxation by annular FIB milling ....................................... 135

8.4 Discussion ......................................................................................................... 140

Microstructure in the internal oxidation zone ............................................ 140

Strain and stress in the internal oxidation zone ......................................... 141

Stress relaxation in the IOZ during oxidation ............................................ 144

8.5 Conclusion ......................................................................................................... 145

9 Conclusion ................................................................................................................ 151

10 Further work.......................................................................................................... 155

xv

List of abbreviations and symbols

Units

° Degree

A Ampere

Å Ångström

at.% Atomic percent

h (and hr) Hour

HV Vickers hardness

Ks-1 Kelvin per second

min minute

Pa Pascal

Pas.s Pascal-second

V Volt

wt.% Weight percent

Physical constants

R Universal gas constant

N0 Avogadro number

Variable parameters

𝑅𝑐 Critical cooling rate

𝑇𝑔 Glass transition temperature

𝑇𝑙 Liquidus temperature

𝑇𝑚 Melting temperature

𝑇𝑟 Reduced temperature (T/Tm)

𝑇𝑟𝑔 Reduced glass transition temperature

𝑇𝑥 Onset crystallization temperature

𝑠1 and 𝑠2 X-ray elastic constants

𝑡𝑚𝑎𝑥 Maximum producible thickness of amorphous alloy

∆Gf Gibb’s free energy

∆Hf Enthalpy

∆Hrelax Enthalpy of relaxation

∆Sf Entropy

∆Tx The difference between crystallization and glass transition temperatures

dhkl Interplanar lattice spacing (Crystallography)

dHV Average diagonal length of a Vickers indent

dp Plastic zone size

ℎ FIB milling depth/Pillar height

ℎ, 𝑘, 𝑙 Miller indices of crystal planes

I Rate of nucleation of crystalline nuclei

Kc Fracture toughness

n Natural number

No. Number

P Applied force (hardness measurement)

pO2 Partial pressure of oxygen

xvi

U Rate of growth of crystalline phase

X, Y 2-dimensional coordinates

ε// In plane strain

ηBSE Backscatter coefficient

μs Shear moduli

σ// In-plane stress

σy Yield strength

𝐵 Bulk moduli

𝐷 Cylinder (or pillar) size/diameter

𝐸 Young’s modulus

𝑆 Solidity index

𝑇 Temperature

𝑉 Volume per unit mass

𝑏 Shape factor

𝑐 Single crystal elastic coefficient (XRD)

𝑓 Fraction of nucleus sites at the growth interface

𝑛 Surface normal (XRD)

𝑟 Atomic radius

𝑡 Time

𝛼 Dimensionless parameter related to the liquid/solid interfacial energy

𝛽 Dimensionless parameter related to the liquid/solid interfacial energy

𝜂 Viscosity

𝜃 Diffraction semi-angle

𝜆 Wavelength (XRD)

𝜇 Linear absorption coefficient (XRD)

𝜈 Poisson ratio

𝜏 Information depth (XRD)

𝜓 Tilt angle

𝜙 Rotation angle

Abbreviations

2D 2-dimensional

3D 3-dimensional

AM Additive manufacturing

BAA Bulk amorphous alloy

BF Bright field

BMG Bulk metallic glass

BMGC Bulk metallic glass composite

BSE Back-scattered electron

CCD Charged coupled device

CCT Continuous cooling transformation

c-Cu2O Cubic cuprous oxide / copper (I) oxide

CEN Center for Electron Nanoscopy at DTU

CTE Coefficient of thermal expansion

c-ZrO2 Cubic Zirconia

DF Dark field

DIC Digital image correlation

DK Denmark

DP Diffraction pattern

xvii

DTA Differential thermal analyzer

DTU Technical University of Denmark

EDS Energy dispersive X-ray spectroscopy

EMC Electron Microscopy Centre at UOW

FEG Field emission gun

FIB Focused ion beam

GFA Glass forming ability

IOZ Inner oxidation zone

JCPDS Joint committee on powder diffraction standard

m-CuO Monoclinic cupric oxide / copper (II) oxide

MEK Department of Mechanical Engineering at DTU

MG Metallic glass

MTU Section of Materials and Surface Engineering at DTU

NTE Negative thermal expansion

OOZ Outer oxidation zone

RS Residual stresses

RSP Rapid solidification process

SAED Selected area electron diffraction

SE Secondary electron

SEC Single elastic constant

SEM Scanning electron microscope/microscopy

SLR Super-cooled liquid region

SMAT Surface mechanical attrition treatment

TEM Transmission electron microscope/microscopy

TGA Thermogravimetry analyzer

TLD Through the lens detector

TPF Thermoplastic forming

TTT Temperature-Time-Transformation

t-ZrO2 Tetragonal Zirconia

UOW University of Wollongong

XEC X-ray elastic constant

XRD X-ray diffraction

Introduction

1

1 Introduction

1.1 Research framework and hypothesis

Generally, metallic materials (alloys) are composed of coherent crystalline grains forming

(mono- and poly-) crystals. In each crystalline phase, the atoms are arranged in a periodic

manner with a well-defined lattice structure. The crystals contain crystalline imperfections

like foreign atoms, dislocations, grain (and phase) boundaries, and inclusions. These

imperfections determine to a high degree the properties of a metallic material like their

mechanical (and chemical) response in the service exposure and can be the location where

failure occurs. As opposed to crystalline metallic materials also amorphous metals (metallic

glasses) can be synthesized. This class of materials possesses a disordered atomic-scale

structure, as the composing atoms lack translational periodic arrangement as in a crystal

[1].

The first metallic glass was synthesized in 1960 by rapid solidification of a liquid Au-Si

alloy with a cooling rate of 106 Ks-1. For this alloy with no mixing in the solid-state and a

large difference in the atomic size of Au and Si, the high solidification rate (105-106 Ks-1)

prevents atoms to solidify in their crystalline structure [2]. Only very thin foils of droplets

can be made for such a high cooling rate. More than half-century of research made it

possible to define criteria for predicting the potential of an alloying system to retain a glassy

state even at low cooling rates. In the late 1980s, Akahisa Inoue proposed the confusion

principle empirically according to which an amorphous alloy can be formed at an

appreciably lower cooling rate (10-103 Ks-1), thus allowing a minimum diameter of 1 mm,

the so-called bulk metallic glass (BMG) [3]. According to Inoue’s criteria, the number of

elemental constituents with a size disparity and a negative heat of mixing (∆Hmix), tends to

suppress nucleation and growth of the crystalline phases during solidification. BMGs have

been developed in several multicomponent alloying systems [4].

Glassy materials, irrespective of whether they are oxidic, metallic or polymeric, are defined

by their glass transition temperature (Tg), i.e. the temperature at which the viscosity of the

undercooled liquid becomes extremely high and the disordered atomic arrangement

containing excess free volume is “frozen-in” [3].

Research framework and hypothesis

2

The non-crystalline nature of BMGs makes them attractive materials with outstanding

properties: high strength and hardness, high elastic limit, low Young’s modulus, super-

plasticity in the super-cooled liquid region (SLR) and, in some cases, extremely good

corrosion and wear resistance. This combination of properties allows BMGs to be exploited

in several engineering applications [5,6]. Despite having exceptional mechanical properties

as high yield strength and high elastic limit among all metallic materials, at the same time,

they exhibit poor ability to deform plastically compared to (poly-) crystalline metals, in

particular under tensile loading, which consequently restricts their applicability [7].

However, high ductility (and toughness) is an essential complementary requirement for an

engineering material in order to be used in structural applications. Generally, crystalline

metals are capable of exhibiting a work-hardening 1 effect and sustaining plastic

deformation. Conversely, in glassy metallic materials, plastic deformation is accompanied

by strain-softening, so combining high strength and high ductility for a material recognized

by its inherent brittleness is technically a major challenge. To overcome this, a series of

methods have been suggested to circumvent the intrinsic brittleness of glassy metallic

materials. In principle, these strategies focus on the development of techniques to

effectively mitigate the propagation of shear bands and thereby improve the resistance of a

BMG against fracture. These methods are surveyed and reported by many authors [8–10].

Technologically, surface modification treatments, among other approaches, can also be

exploited to enhance the plasticity (and ductility) of metallic glasses. Ref. [11] reviews

some of the ideas to use surface treatment as an effective method to increase the toughness

of metallic glasses by the introduction of compressive residual stresses in their surface

region to prevent crack nucleation.

Among various surface treatments, low-temperature gaseous thermochemical treatment is

widely used to enhance the performance of alloys and metals against surface-initiated

failure mechanisms, such as fatigue, wear, and corrosion [12]. Gaseous thermochemical

treatment entails deliberately changing the surface composition at elevated temperatures in

1 The increase of stress with strain in the plastic region of a stress–strain curve is known as

work-hardening (strain-hardening, as the stress necessary for continued straining increases

with the total strain).

Introduction

3

a controlled atmosphere to achieve a hardened case, typically by the incorporation of

interstitial elements like oxygen. The starting point for the present Ph.D. project was the

hypothesis that surface engineering by thermochemical post-processing of a BMG in

oxygen-containing atmospheres can be applied as an effective method to enhance the wear

resistance and to introduce compressive residual stresses in the surface region. Moreover,

it is anticipated that the build-up of residual stresses as a consequence of an oxygen-

diffusion in the surface of a BMG could play a significant role in the development of the

surface-oxidation zones.

To this end, the overall hypothesis of this project is that:

- There is a largely unexplored potential to improve the surface hardness and

introducing compressive residual stresses in the surface region of BMGs employing

gaseous thermochemical oxidizing treatment, which would allow a wider application

of these materials where “peculiar” (combinations of) properties are necessary.

- The ductility of BMGs can be improved through surface engineering, by introducing

compressive residual stress in the surface region such that the crack initiation is

effectively postponed to higher applied stresses.

- Understanding the correlation between the oxidizing treatment parameters,

microstructure evolution, and the development of the residual stresses in the surface

zones of the oxidized BMGs.

1.2 Motivation and research objectives

The present Ph.D. project is systematic research financially funded by “Villum Fonden”,

Denmark2. The proposed project seeks to explore the oxidation mechanism of the multi-

component noble-metal containing Zr-based BMGs as well as determining how the residual

stresses develop in the hardened oxidized regions using various ex-situ and in-situ

techniques. The investigations contribute to the fundamental knowledge of how oxidizing

in different oxygen partial pressure (pO2) atmospheres influences the development of

2 VILLUM FONDEN is a non-profit private foundation that supports excellent technical

and scientific research as well as environmental, social and cultural purposes in Denmark

and abroad.

Scientific and experimental approach

4

crystalline oxide zones in the surface adjacent region of the BMGs. Emphasis is on

reaching an understanding of the relation between the oxidizing conditions, surface (micro)

structural changes, and residual stress evolution using advanced microstructure

characterization methods, which will be of interest from both scientific and technological

points of view.

1.3 Scientific and experimental approach

This Ph.D. project is meant to establish knowledge-based innovative solutions and

concepts. The study departs from a systematic investigation of various Zr-based BMGs to

screen their surface response to oxidizing. Only fragmentary information can be found in

the literature, illustrating that in principle the room temperature plasticity of BMGs can be

improved by shot peening [13], which introduces compressive residual stresses in the

surface that counteract stress concentration and retard crack initiation. A related mechanical

surface engineering technique is surface mechanical attrition treatment (SMAT) [14].

Optimizing the SMAT processing parameters, the plastic strain can be increased.

Controlled (thermal) surface crystallization leads to the development of isolated crystallite

islands at the surface, which may act as obstacles for the propagation of shear bands. Lastly,

an electrodeposited coating to encapsulate BMGs was demonstrated to have a beneficial

effect [15]. Among these surface modification techniques, surface engineering by gaseous

thermochemical treatment has received practically no attention. Therefore, in the present

Ph.D. project, the incorporation of the dissolved elements like oxygen in the surface of the

BMG is investigated.

The investigations focus on three Be-free ZrCuAl-based BMGs, which are known as good

glass-forming alloys. At the same time, these BMGs are prone to oxidize, because of the

presence of strong oxide-forming elements like Zr and Al, which both have a high affinity

for oxygen. In the experimental procedure, after applying some initial characterization on

selected BMGs with X-ray diffraction and thermal analysis, gaseous (thermochemical)

oxidizing treatment is performed in a different oxidizing atmosphere, aiming at determining

how the oxygen dissolution in the glassy metal influences the development of surface-

oxidation zones.

For this purpose, the oxidation is followed in-situ and ex-situ applying thermo-gravimetry

and X-Ray diffraction. Characterization of the surface-oxidation structure is performed ex-

Introduction

5

situ (post-mortem) using various light and electron microscopy techniques. Angular

dispersive X-ray diffractometry was applied, using Cu (or Cr) radiation to investigate ex-

situ (and in-situ) the crystalline oxide phases developing in the surface-adjacent region of

the oxidizing BMGs. Lastly, (lattice) strains and associated residual stresses in the surface-

oxidized BMGs were determined using conventional ex-situ (and in-situ) X-ray based sin2ψ

method and incremental ring-core focused ion beam milling (FIB) together with digital

image correlation (DIC) methods.

1.4 Outline

The present doctoral thesis is structured in 10 individual chapters (see Fig. 1.1). Chapter 1

introduces the background, research motivations, and provides an overall scientific and

experimental approach of the Ph.D. work. In chapter 2, relevant literature on BMGs, their

functional properties, and techniques employed to enhance their ductility (and toughness)

are reviewed. This chapter also gives the general features of the oxidation behavior of

metallic glasses with an emphasis on Zr-based amorphous alloys. In chapter 3, the

experimental procedure and a brief overview of the (characterization) techniques utilized

in this study are addressed. A summary of the main research findings provided in the results

chapters is presented in chapter 4. The result chapters (5-8) are in the form of manuscripts

which provide detailed descriptions and comprehensive discussions of the experimental

results. The Ph.D. dissertation is completed with the conclusions (chapter 9) followed by a

remark on future works (chapter 10).

Fig. 1.1. Flowchart on the structure of the Ph.D. dissertation.

Outline

6

References

[1] M.K. Miller, P. Liaw, Bulk Metallic Glasses: An Overview, 2007.

[2] W. Klement, R.H. Willens, P. Duwez, Non-crystalline structure in solidified gold-

silicon alloys, Nat. Sci. 187 (1960) 869–870.

[3] C. Suryanarayana, A. Inoue, Bulk Metallic Glasses, CRC Press, 2011.

[4] A. Takeuchi, A. Inoue, Classification of Bulk Metallic Glasses by Atomic Size

Difference, Heat of Mixing and Period of Constituent Elements and Its Application

to Characterization of the Main Alloying Element, Mater. Trans. 46 (2005) 2817–

2829.

[5] A.L. Greer, E. Ma, Bulk Metallic Glasses : At the Cutting Edge of metals research,

MRS Bull. 32 (2007) 611–619.

[6] M.F. Ashby, A.L. Greer, Metallic glasses as structural materials, Scr. Mater. 54

(2006) 321–326.

[7] R.O. Ritchie, The conflicts between strength and toughness, Nat. Mater. 10 (2011)

817–822.

[8] Y. Sun, A. Concustell, A.L. Greer, Thermomechanical processing of metallic

glasses: Extending the range of the glassy state, Nat. Rev. Mater. (2016).

[9] D.C. Hofmann, J.-Y. Suh, A. Wiest, G. Duan, M.-L. Lind, M.D. Demetriou, W.L.

Johnson, Designing metallic glass matrix composites with high toughness and

tensile ductility, Nature. 451 (2008) 1085–1089.

[10] S. V. Madge, D. V. Louzguine-Luzgin, J.J. Lewandowski, A.L. Greer, Toughness,

extrinsic effects and Poisson’s ratio of bulk metallic glasses, Acta Mater. 60 (2012)

4800–4809.

[11] T.G. Nieh, Y. Yang, J. Lu, C.T. Liu, Effect of surface modifications on shear

banding and plasticity in metallic glasses: An overview, Prog. Nat. Sci. Mater. Int.

22 (2012) 355–363.

[12] E.J. Mittemeijer, M.A.J. Somers, Thermochemical Surface Engineering of Steels,

Woodhead Publishing Limited, 2015.

[13] Y. Zhang, W.H. Wang, A.L. Greer, Making metallic glasses plastic by control of

residual stress, 5 (2006) 857–860.

[14] J. Fan, A. Chen, J. Wang, J. Shen, J. Lu, Improved plasticity and fracture toughness

in metallic glasses via surface crystallization, Intermetallics. 19 (2011) 1420–1427.

[15] L.W. Ren, M.M. Meng, Z. Wang, F.Q. Yang, H.J. Yang, T. Zhang, J.W. Qiao,

Enhancement of plasticity in Zr-based bulk metallic glasses electroplated with

copper coatings, Intermetallics. 57 (2015) 121–126.

Background

7

2 Background

The introductory chapter includes the concept of (metallic) glass formation, alloying

design, and scientific literature and progress in the field of bulk metallic glasses (BMGs).

The focus is on fabrication methods, functional material properties as well as application

fields of BMGs, techniques developed for enhancing ductility (and toughness) of BMGs

including surface modification treatments, and studies of the oxidation behavior of

amorphous alloys mostly Zr-based BMGs.

2.1 Metallic Glasses (MGs)

The term “glass” is associated with any inorganic products of fusion, i.e. liquid1 that have

been cooled continuously to form a condensed solid2 without crystallizing. To elucidate the

general properties of glass, an improved definition for glass was recently proposed by E.

D. Zanotto and J. C. Mauro [1]:

Glass is a nonequilibrium, non-crystalline condensed state of matter that exhibits a glass

transition. The structure of glasses is similar to that of their parent supercooled liquids3,

and they spontaneously relax toward the supercooled liquid state. Their ultimate fate, in

the limit of infinite time, is to crystallize.

In addition to oxide glasses (like silica known as hard, brittle, and transparent substance),

polymers, and more recently hybrid metal-organic frameworks [2], metals can also turn to

glass. In common practice, most metallic alloys crystallize during solidification, arranging

their atoms to a highly-ordered structure. On the other hand, if crystallization does not occur

1 A liquid is a fluid, condensed state of matter that exhibits viscous flow [1].

2 The solid is a state of condensed matter where the atomic structure is thermodynamically

stable and the chemical forces are strong enough to keep the structure cohesive. The solid

can be crystalline or amorphous [1].

3 Supercooled liquids are thermodynamically metastable liquids that exist between the

melting point and the glass transition temperature [1].

Metallic Glasses (MGs)

8

during cooling of liquid metal to room temperature, a closely-packed low-density

(compared to crystalline solids due to the presence of excess free volume) disordered

material lacking the periodicity of crystals is obtained, i.e. metallic glass [3].

The concept of glass formation

The formation of glass is achieved by continuous cooling from the liquid state. As presented

in Fig. 2.1-a, on cooling a molten metal, the specific volume (volume per unit mass, V) and

the enthalpy gradually decrease as the temperature (T) approaches its melting point, Tm. At

Tm, the volume of the liquid either decreases sharply on transforming into a crystalline solid

or continues to decrease at the same rate, and becomes super-cooled liquid (SCL). The

former takes place by nucleation and growth of the crystalline phases at relatively slow

solidification rates and the crystalline (long-range ordered) solid contracts with decreasing

temperature. The latter occurs at relatively high cooling rates without crystallization at Tm.

The volume decreases continuously in the super-cooled region while the viscosity, i.e. the

resistance of a fluid against flow under applied shear stress, increases. At a certain

temperature far below Tm, the specific volume experiences a change in slope and continues

almost parallel to the contraction of the crystal. At this temperature the viscosity of the

super-cooled liquid has become so high (~ 1012 Pa.s) [4], the atoms freeze in their

configuration with respect to each other as a non-equilibrium “disordered frozen-in” state:

a glass. In other words, the atoms do not have enough time to arrange themselves to the

metastable equilibrium, and thereby the structural arrangement is kinetically arrested

leading to the deviation of volume (or enthalpy) line of SCL [5]. The temperature at which

the super-cooled liquid becomes a glassy solid is termed as glass transition temperature, Tg.

Here it is noted that unlike Tm, which is a thermodynamically discrete temperature, Tg is

kinetically dependent and increases with increasing cooling/heating rates. This means that

since the glass transition is a kinetic freezing phenomenon, the final volume (or enthalpy)

of glass be contingent on cooling rates; when the cooling rate is high, the system has

insufficient time to relax into a lower energy state leading to a higher enthalpy (more excess

free volume) in a final glass [5]. In addition to the V-T graph, the variation of viscosity (η)

with temperature during cooling from the liquid state is provided in Fig. 2.1-b. On cooling,

the η of a molten metal increases slowly with a decrease in temperature. At Tm, the viscosity

of the material increases abruptly by about 15 orders of magnitude to form a crystal.

However, the variation of η in a glass-forming liquid is different. Here, below the melting

Background

9

point η increases gradually with decreasing temperature in the undercooled liquid region.

At Tg, the η becomes so high which leads to no more flow of the liquid and thereby the

material develops into a solid glassy metal [4].

Fig. 2.1. (a) Variation of specific volume with temperature for a crystal and a glass-forming material. (b) The

viscosity (η) with temperature for crystal and glass formation. Reprinted and recreated from ref. [4].

As mentioned, the nucleation and growth of the crystalline phases during cooling of a

molten alloy can be kinetically by-passed at relatively high4 solidification rates yielding a

frozen liquid, i.e. metallic glass. The ability of a metallic alloy for this transformation is

defined as glass-forming ability (GFA). One of the key points to evaluate the GFA is the

critical cooling rate (Rc) required for transformation into a glassy state. The determination

of an Rc can be easily described with regard to Fig. 2.2 which represents a schematic

temperature-time-transformation (TTT) 5 diagram for an alloy [4]. The C-shape

transformation curve indicates the time required for the formation of a crystalline phase at

4 A relatively high solidification rate is with respect to the redistribution of the atoms (or in

oxidic and polymeric materials the molecules) in the liquid.

5 TTT diagram is generally used to illustrate the idea of obtaining either crystalline or glassy

state from the liquid, but this is technically incorrect. TTT diagrams show the time required

for a phase transformation at constant temperature (isothermal behaviour). Here, the

continuous cooling transformation (CCT) from the liquid and the effect of cooling rates on

the transformation are presented.


10

any given temperature. When the molten alloy is cooled from above the liquidus

temperature (Tl), at a rate shown as path “1”, solidification starts at Tm and time tm, and

thereby a crystalline solid is obtained. However, solidifying the same molten alloy at a

higher cooling rate indicated by path “2”, the liquid becomes “frozen-in” and a glassy phase

is formed below Tg. The cooling rate indicated by path “2” is the Rc required for the glass

formation. If the corresponding temperature and time at the knee of the C-curve are Tn and

tn, respectively, the Rc is given by l nc

n

T TR

t

[4]. When the liquid-solid transformation (C-

shape curve) is shifted to the right, it suggests that the liquid can be retained longer in the

super-cooled condition and therefore the value of Rc for the glass formation is lower. Further

the C-curve is shifted to the right, the lower is the value of Rc and the higher value of GFA

is obtained. In other words, a metallic with a high GFA can be synthesized in a larger

critical dimension (or thickness) at lower Rc (cf. Fig. 2.3). In addition to Rc, i.e. an effective

indicator of the GFA of the alloy melt, several other parameters have been proposed to

assess the GFA. The first criterion for determining the GFA, reported by Turnbull [6], is the

reduced glass transition temperature (g

rg

l

TT

T ).

The ratio g

l

T

T arises from the fact that the viscosity of the super-cooled liquid should be

high at a temperature between the melting point and the glass transition for the formation

of a glass. The viscosity at Tg is constant, thus the higher the ratiog

l

T

T, the higher the

viscosity at the knee of the TTT (or CCT) diagram, and thereby results in the lower Rc for

glass formation [7]. Moreover, it is empirically proven that changing the chemical

composition of the alloy has a negligible effect on Tg value, but on the other hand, Tl is

greatly influenced by changing the elemental constituent concentration. Therefore,

designing the alloying system exhibiting lower Tl, results in enhancing the probability of

being able to solidify through the interval between Tl and Tg without devitrification, i.e. the

improvement of GFA [7].

Background

11

Fig. 2.2. Schematic illustration of TTT diagram of a liquid-solid transformation for a hypothetical alloy.

Reprinted and recreated from ref. [4].

It has been stated that the devitrification of the alloying systems having Trg ≥ 2/3, can

completely be suppressed. Moreover, another representative GFA criterion proposed by

Inoue is the super-cooled liquid region (SLR), i.e. the temperature difference between the

onset crystallization temperature Tx and the Tg (ΔTx =Tx-Tg) [8]. It has been shown that

alloying systems with a large ΔTx, possess a high GFA. From the empirical relations among

Rc, Trg, tmax (maximum producible thickness), and ΔTx, several multicomponent amorphous

alloys were successfully synthesized indicating good GFA where some of them are

summarized in Fig. 2.3.

Fig. 2.3. The empirical relations between the critical cooling rate for glass formation (Rc), maximum sample

thickness (tmax), and the reduced glass transition temperature (Trg=Tg/Tm) or the super-cooled liquid region

(∆Tx=Tx-Tg) for some amorphous alloys [9].


12

Development of bulk metallic glasses (BMGs)

The size of the early metallic glasses fabricated by rapid melt quenching process (105-106

Ks-1) is constrained to a few tens of micrometers thick due to the high Rc required for the

glass formation. Therefore, they can only be utilized in a limited number of applications as

for instance iron-based amorphous ribbons used as soft-magnetic material [10]. To improve

their applicability, attention has turned to the development of multicomponent glassy

alloying systems allowing the synthesis of metallic glasses in thicker sections and at lower

cooling rates. These research efforts were successful in 1974 when Chen et al. [11] reported

the fabrication of 1-3 mm Pt–Ni–P and Pd–Ni–P metallic glasses using water quenching.

In 1984, Kui et al. [12] utilized a fluxing technique for suppressing heterogeneous

nucleation during the solidification process using B2O3, which consequently lead to the

fabrication of 10 mm Pd40Ni40P20 BMG. These observations were subsequently confirmed

by Peker and Johnson [13] at Caltech where they fabricated a 14 mm diameter fully glassy

rod with nominal composition Zr41.2Ti13.8Cu12.5Ni10.0Be22.5 (known as Vitreloy 1). Based on

the presence of multi-alloying constituents in the chemical composition of BMGs as some

of them are provided in Fig. 2.3, Inoue proposed three empirical criteria (confusion

principle) for slow crystallization kinetics and thus synthesizing BMGs having a stabilized

super-cooled liquid state and high GFA [4,9]:

The existence of at least 3 main constituent elements in the chemical composition

of the alloy to increase the complexity of the system, i.e. entropy of the system.

The difference of the atomic radius between the main alloying element and the

other constituents should be greater than 12%.

The heat of mixing enthalpy should be negative between the main elements.

In addition to the above-mentioned criteria, it has also been reported that choosing the

chemical composition of the alloy adjacent to the deep eutectic composition would be

beneficial for increasing the ability of the alloy melt for glass formation [14]. These criteria

are empirically obtained based on the (i) thermodynamic, (ii) kinetic, and (iii) structural

aspects of a glass formation. Thermodynamically speaking, Gibbs energy ∆Gf (T)=∆Hf -

T∆Sf for the transformation of a liquid to a glassy state should be lower compared with the

formation of the crystalline phase(s). In other words, a glass becomes more “stable” when

the Gibbs energy of the glass phase is lower compared with the competing crystalline

phases, i.e., ∆Gf=Gglass-Gcrystal becomes negative [4]. This Gibbs energy reduction could

Background

13

be obtained by increasing the entropy of the system, ∆Sf, or by decreasing the enthalpy of

mixing, (negative) ∆Hf, preferably the contributions of both factors. The value of ∆Sf is

expected to increase by the multiplicity of alloy components, which results in an increase

in the random atomic distribution, as well as an increase of the energy of the solid/liquid

interface. Moreover, these criteria for the formation of high GFA glassy alloy are important

from the kinetic point of view. The rate of homogenous nucleation, I and growth, U of a

spherical crystalline phase from the super-cooled liquid are expressed by the following

equations [9,15,16]:

30 3 210 / .exp[ / ( .(1 ) ]r rI b T T [cm-3s-1] 2-1

210 / .[1 exp[ . / .( / )]]r r mU f T T T T [cms-1] 2-2

where Tr is the reduced temperature (T/Tm), ∆Tr is the difference in temperature from Tm, b

is a shape factor (e.g., 16π/3 for a spherical nucleus), η is viscosity and f is the fraction of

nucleus sites at the growth interface, α and β are dimensionless parameters related to the

liquid/solid interfacial energy (σ); α = (N0V)1/3.(σ/∆Hf) and β= ∆Sf /R. Here N0, V, and R

are Avogadro number, atomic volume, and gas constant, respectively. Apparently, the

increase in η, α, and β parameters, decreases the nucleation and growth rate of the formation

of crystalline phases, and thereby improved GFA is achieved. The decrease in ∆Hf and

increase in ∆Sf are consistent with obtaining a better GFA from the thermodynamics, too.

Moreover, a significant atomic radius mismatch (∆r/r) between the alloying elements

(≥12%) is advantageous for the formation of a high-dense random atomic structure due to

the difficulty of atomic rearrangements to form a crystal.

Based on the confusion principle (∆Hf and ∆r/r), BMGs are classified into seven groups

presented in Fig. 2.4. These seven groups include: (G-I) ETM/Ln-LTM/BM Al/Ga, (G-II)

ETM/Ln-LTM/BM-Metalloid, (G-III) Al/Ga-LTM/BM- Metalloid, (G-IV) IIA-ETM/Ln-

LTM/BM, (G-V) LTM/BM-Metalloid, (G-VI) ETM/Ln-LTM/BM and (G-VII) IIA-

LTM/BM, where ETM, Ln, LTM, BM, and IIA refer to early transition, lanthanide, late

transition, late transition, IIIB-IVB (In, Sn,..) and group alkaline metals in the periodic

table, respectively [17].


14

Fig. 2.4. Classification of BMGs, representing seven groups symbolized by G-I to G-VII. The ETM and LTM

represent the transition metals belonging to groups IIIA–VIIA and VIII–IIB in the periodic table, respectively.

Reprinted and recreated from [17].

Fabrication methods

As mentioned above, a glassy alloy can be synthesized by rapid cooling of a liquid to avoid

nucleation and growth of the crystalline phases. Therefore methods that allow a fast

extraction of heat from the liquid are utilized [4,18].

Various techniques have been exploited to achieve a high solidification rate, such as gas

atomization, twin roller quenching, and melt-spinning methods (see Fig. 2.5). In the gas

atomization method, the flowing molten metal from a nozzle is impinged by a cold stream

of air or an inert gas (N2, He, and Ar). Then the atomized molten droplets solidify as powder

on a thermally conductive substrate. In the twin roller quenching method, a stream of

molten metal solidifies as a continuous filament, ribbon, or sheet in contact with two

(copper) chilling cylinders as presented schematically in Fig. 2.5-b. Another method used

for achieving high solidification rates is the melt-spinning process. In this technique,

molten metal is led to a single fast-rotating internally cooled disk. The molten metal stream

rapidly solidifies in contact with the surface of the cooled disk (heat-extracting surface),

allowing the continuous fabrication of an amorphous ribbon/tape (cf. Fig. 2.5-c) [19,20].

Background

15

Fig. 2.5. Various techniques for rapid solidification process (a) gas atomization method. Reproduced and

redesigned from [21] , (b) twin roller quenching method. Reproduced and redesigned from [22], and (c) melt-

spinning method [23].

The criteria of glass formation in bulk form (above 1 mm in diameter or thickness) at low

cooling rates (10-103 Ks-1) causes developing several casting techniques to prepare BMGs

during the past decades. Up to now, extensive research activities resulted in the fabrication

of Pd- [24], Zr- [25], Ti- [26], Fe- [27], Co- [28] and Cu-based [29] BMGs using copper-

mold, suction, high pressure die [30], and cap casting [31] methods as well as powder

metallurgy [32].

These commonly used processes enable the manufacturing of MGs in simple geometries

and small sizes. In general, synthesizing more complex and larger components is beyond

the capabilities of the mentioned methods due to the restricted GFA of the amorphous

alloys. In addition to these techniques, the superplastic nature of the amorphous metallic

alloys within the SLR (∆Tx=Tg-Tx), allows thermoplastic forming (TPF) of the BMGs [33].

This method could be utilized to form BMGs in several geometries; however, the

processing window, i.e. the required time for TPF is restricted to prevent devitrification of

a glassy alloy [34]. In other words, near-net-shape of the sophisticated geometries can only

be achieved for good glass formers having higher ∆Tx without crystallization of the bulk.

Nevertheless, most of the BMGs can only be synthesized in parts with diameters less than

20-30 mm, which is a great challenge in finding the right application for these small parts.

Therefore, to overcome the limitation in the fabrication of BMGs in terms of size and

complicated geometries as well as having an optimum combination of properties, recently,

additive manufacturing technology (AM) has been used for synthesizing BMGs. This

method allows rapid prototyping and fabricating of complex structural parts directly from

CAD design [35–38]. The three-dimensional complicated BMG components have been


16

fabricated by progressively adding thin layers of metallic glass powders guided by a 3D

digital model using the powder bed fusion (PBF) technique.

Moreover, the fabricating of bulk metallic glass composite (BMGC) has received attention

in recent years as a potential method to enhance some of the properties of BMGs in service

exposure via the introduction of micro or nano-size second phase crystalline phases in an

amorphous matrix [39–42]. BMGC fabrication is successfully achieved either by in-situ or

ex-situ processes [43]. In an in-situ metallic glass composite preparation, the chemical

composition of the alloying system is designed in a way that during solidification, initially

second crystalline phase forms (nano-crystallites and dendrites) and then the remaining

melt solidifies as glass. However, in an ex-situ BMGC fabrication, the second phase is

added directly to the pre-alloyed metallic glass melt during casting. The second phase

varies from particles to short (and long) fibers within the amorphous matrix. The type of

reinforced agent used for making metallic glass composites were pure metals (W, Ta, etc.)

and ceramics (SiC, TiC, WC, etc.). These embedded inhomogeneities in the amorphous

matrix serve as structural blocking sites for shear band propagation and mitigate the crack

nucleation [44,45]. The distribution, shape, size, and volume fraction of the precipitates can

play an effective role in the mechanical response of the final products. This could be

achieved by controlling several technical parameters such as the solidification rate, partial

crystallization, and chemical composition of the alloying system.

Functional properties of (bulk) metallic glasses

The lack of an ordered atomic arrangement in the glassy metals results in the absence of

crystalline defects such as grain (and phase) boundaries and dislocations in their atomic

structure. This makes them attractive engineering materials with a unique combination of

mechanical, chemical, and physical properties including high strength and hardness, low

Young’s modulus, high elastic limit, in some cases good corrosion and wear resistance, and

high magnetic permeability (Co and Fe-based MGs). As presented in Fig. 2.6 a-b, which

shows the variation of tensile strength and hardness versus Young’s modulus of some

crystalline and glassy alloys, BMGs have superior strength and hardness as compared to

conventional crystalline alloys. Comparing Young’s modulus (E) of crystalline and BMG

alloys shows that BMGs have lower E values. Therefore, a combination of high yield

strength and low Young’s modulus of BMGs exhibits considerably higher elastic limit of

Background

17

these defect-free materials as compared to conventional metallic alloys, suggesting that

they would be suitable for applications that require elastic materials [46,47].

Fig. 2.6. The variation of (a) tensile strength and (b) Vickers hardness against Young’s modulus of some

metals and BMGs [4]. (c) Ashby map showing the fracture toughness versus yield strength of various

materials. Diagonal lines indicate the plastic-zone size, K2c/πσ2

y, where Kc is the fracture toughness and σy is

the yield strength [48].

Fig. 2.6-c shows an Ashy map (materials-selection map) comparing fracture toughness

(Kc)6 against yield strength (σy)

7 of several engineering materials including oxide glasses,

ceramics, polymers and metals together with data for the Pd79Ag3.5P6Si9.5Ge2 BMG shown

by a star, values for other BMGs (including three Fe-based; two Zr-based; a Ti-based, and

6 The resistance of a material to cracking is characterized by either fracture toughness, i.e.

the maximum load (stress) that a material can withstand before fracture or toughness, i.e.

the critical energy absorbed by the matter before rupture [46].

7 In material science and engineering, yield strength is defined as maximum stress at which

an object deforms elastically before it begins to change its shape (plastic deformation). In

other words, the yielding indicates the limit of elasticity of a material and its plastic

behavior.


18

a Pt-based BMGs) marked by crosses and data for ductile-phase-reinforced BMG

composites presented by circles [48]. As shown in the map, conventional oxide glasses

exhibit low impact resistance and are considered as brittle materials. Moreover, most

BMGs show limited fracture toughness (except Pd79Ag3.5P6Si9.5Ge2 BMG exhibiting

highest damage tolerance among engineering materials [49] and most of BMGCs [50]),

where their fracture toughness-strength relationship lies between the brittle ceramics and

ductile metals. Fig. 2.6-c also presents contour lines indicating the value of plastic zone

size,2

2

cp

y

Kd

, where Kc is the fracture toughness and σy is the yield strength. At the tip of

a sharp crack in a material, plastic flow can occur and blunt the crack. Hence, for a material

with a lower yield stress σy, the plastic flow can facilitate (for example quenched and

tempered steels), and a as result an increase in toughness is obtained [46]. If “dp” is much

smaller than the sample size, crack propagation leading to the occurrence of fast fracture.

On the other hand, if “dp” is greater than the sample dimension, brittle fracture is not

expected. As already stated, BMGs exhibits high yield strength, and thus even those with

high fracture toughness reveals small process-zone sizes. Ideally, the BMGs shows a “dp”

of a few millimeters, compared to approximately 1 m in structural steels. The most brittle

MGs (for example Fe40Ni40P14B6) have smaller “dp” values compared with those for oxide

glasses [46,47].

In addition to their extraordinary combination of mechanical properties, BMGs have good

corrosion resistance and an improved passivation ability due to their ideal single-phase

homogeneous nature. However, some experimental studies revealed limited corrosion

stability of different BMGs depending on their chemical composition and the corrosive

environment [51,52].

Application fields of (bulk) metallic glasses

As mentioned in section 2.1.4, the notable properties of BMGs have qualified them to be

utilized in particular engineering applications. Based on BMGs’ fundamental

characteristics, some of their current commercialized applications and some possible

applications for future technologies are summarized in table 2.1.

Background

19

Table. 2.1. Current and possible applications of BMGs [4,14,53,54].

Application Technical reasons Example BMG alloy

Sport equipment High strength and elastic limit Golf clubs, tennis

racquets, baseball bats Zr55Cu30Al10Ni5

Medical devices High hardness, high wear,

and corrosion resistance

Ophthalmic scalpel

blades, minimally

invasive medical

devices, medical

stapling anvils

ZrCuAlAgSi,

Liquidmetal Zr-based

BMG

Edges for sport

goods

High hardness, high wear

resistance Ski, skates, snowboard


BMG

Biomedical

High strength and hardness,

high wear and corrosion

resistance, highly

biocompatible, non-allergenic

element content

Tooth implants, plates,

and screws for fixing

bones

Ti40Zr10Cu36Pd14

Wires High strength and workability Musical instruments Zr-Cu-Ni-Al-Ti

Luxury goods

High hardness, high wear,

and corrosion resistance,

lustre, precious metal content,

shock resistance, light, anti-

magnetic

Jewellery and wrist

watches (case, bezel,

winding crown)

Au49Ag5.5Pd2.3Cu26.9Si16.3,

Pt57.5Cu14.7Ni5.3P22.5, Zr-

Cu-Ni-Al-Ti

High-

performance

spring

High strength, high elastic

limit, high fatigue strength,

and high workability

Automobile valve

spring Zr55Cu30Al10Ni5

Defence Similar density to Uranium

penetrator, Self-sharpening

Anti-tank armour-

piercing projectiles W-reinforced BMGC

Aircraft High strength, light, easy

formability in the SLR

Slat-track cover at the

front of the airplane

wings

Zr55Cu30Al10Ni5

Sensors Low Young’s modulus

Diaphragm for a

pressure sensor, pipes

for a Coriolis mass

flowmeter

Zr55Cu30Al10Ni5

Ti50Cu25Ni15Zr5Sn5

Electronic

devices

High wear resistance, High

wear resistance, thinner

casing with sufficient

mechanical strength, easy

formability in the SLR

Mobile phone and

digital camera casing,

materials for master

discs


BMG

Magnetic High magnetic permeability

and low coercivity

Power distribution

transformers, magnetic

yoke (Fe-based MGs)

Fe73Ga4P11C5B4Si3

Shot peening

high strength, high

endurance against cyclic

bombardment load

Shot peening balls Fe44Co5Ni24Mo2B17Si8

Chemical High corrosion resistance,

easy formability in the SLR Fuel cell separator Ni60Nb15Ti15Zr10

Motors

High strength and hardness,

high wear resistance, high

durability

Micro-geared motors

with high torque

Ni53Nb20Zr8Ti10Co6Cu3,

Zr44Ti11Cu10Ni10Be25

Optical Smooth surface finish, high

reflectivity, no grain structure Optical mirror device Zr55Cu30Al10Ni5

Ductility and toughness enhancement of BMGs

20

Although BMGs have exceptional elastic properties, their plastic deformation at room

temperature is associated with inhomogeneous flow in highly localized shear bands. Shear

banding in BMGs causes failure, which appears macroscopically brittle rather than ductile.

Moreover, BMGs do not exhibit strain-hardening like conventional metals and the

formation of multiple shear bands could cause a limited plastic deformation when subjected

to loads beyond the elastic limit (yield strength) especially in compression (and bending).

Here it is noted that most of the BMGs fail catastrophically with no ductility (or plastic

strain) during tensile loading. Therefore, the lack of ductility (or plastic strain) in BMGs

restricts their applicability in engineering fields. To overcome this softening, strategies

were developed to effectively retard the propagation of shear bands and thereby extend the

resistance of a BMG against (brittle) fracture.

2.2 Ductility and toughness enhancement of BMGs

Unlike crystalline alloys, in which the main deformation mechanism is described by

dislocation motion, plastic deformation in metallic glasses is carried by shear bands

formation [55]. Introducing appropriate techniques for enhancing the ductility and

toughness of BMGs is technically important as regards to their functional performance and

widen their applicability. Previously, some studies have been conducted to improve the

plasticity of metallic glasses via changing their chemical composition to obtain higher

Poisson’s ratio resulting in the reduction of the solidity index of the isotropic material

[56,57] (

3 1 2

2 1

sSB

), where S is the solidity index, ν is the Poisson’s ratio, µs and

B are the shear and bulk moduli, respectively. The equation shows that the intrinsic ductility

increases with increasing ν [58,59]. Yu and Bai [60] evaluated the effect of the Al content

in the bulk glassy (Cu50Zr50)100−xAlx (x = 0–10) system on its compressive plasticity. It was

demonstrated that all the BMG samples owned a similar elastic strain of about 1.7%.

Moreover, they reported that the amorphous samples followed different plastic deformation

behavior with reference to their Al content. The (Cu50Zr50)95Al5 exhibited the maximum

plastic strain of 16% where its Poisson’s ratio is the highest value among the investigated

BMGs. On the other hand, the sample containing 8 at% of Al has the lowest toughness

about 0.5%. The authors supported the notion that the plasticity of the investigated ZrCu-

based BMGs is correlated with their Poisson ratio. Hence, by considering the proper

Background

21

combination of various alloying elements, a ductile metallic glassy system could be

synthesized.

The other method which has formerly been presented for making the metallic glasses

ductile is the fabrication of an amorphous matrix composite entailing crystalline

precipitated second phase [61–63]. The dispersed crystalline particles act as obstacles to

emphatically mitigate the localized deformation of shear bands and their unrestricted

extension [64]. Hofmann et al. [65] investigated the mechanical properties of the three in-

situ BMG composites, Zr36.6Ti31.4Nb7Cu5.9Be19.1, Zr38.3Ti32.9Nb7.3Cu6.2Be15.3, and

Zr39.6Ti33.9Nb7.6Cu6.4Be12.5 where the ratios of Zr, Ti, Nb, and Cu are fixed and the Be

content varies from 12.5-19.1 at.%. They expressed that the body-centered Zr-Nb-Cu

secondary dendritic phases increased by reducing the amount of Be. The results showed

that the total strain to failure was in the range 9.6–13.1% at tensile strengths of 1.2–1.5

GPa. It was shown that the tensile ductility of the investigated BMGCs has noticeably been

improved compared with commercial Vit 1 (Zr41.2Ti13.8Cu12.5Ni10.0Be22.5), indicating no

plastic region in tension. It has been reported that the existence of Zr-Nb-Cu crystalline

second phases resulted in localizing the shear bands around them and suppresses the crack

opening.

Many other investigations suggested that the mechanical properties of BMGs are associated

to their structural stability characterized by the degree of atomic packing density (or free

volume) [66,67]. Therefore, several post-processing treatments have been exploited to

increase the degree of disorder (higher stored energy) within the BMG accompanied by

increased structural defects (free volume) in its atomic structure [68]. These techniques

include plastic deformation, elastostatic loading, and ion (and neutron) irradiation [69–75],

which causes the “rejuvenation” of a glassy metal. This means that an as-cast BMG

undergoes into states having higher stored energy and lower density, generally with

ameliorated mechanical properties. In other words, the abundant free volume facilitates

multiple shear band formation in many local areas of the alloy (decreased localization of

strain) resulting in improved plasticity (and ductility) of a BMG in compression and

tension. The effect of the irreversible structural changes induced by these methods (creation

of excess free volume), can be quantified in terms of an increase in enthalpy of relaxation

(∆Hrelax) of a glass. In addition to this post-treatment, Ketov et al. [76] demonstrated that

the thermal cycling of metallic glasses between room temperature and liquid-nitrogen

Ductility and toughness enhancement of BMGs

22

temperature (77K) can also induce significant structural rejuvenation, reaching less relaxed

states of higher energy in a glass. It was proposed that the thermal cycling of glass is

associated with the development of non-affine internal stresses as a consequence of the

non-uniform coefficient of thermal expansion (CTE). This is because the non-uniform

structure of glass corresponds to local variations in CTE. Cryogenic thermal cycling has

led to a spectacular rejuvenation effect, demonstrating a dramatic increase in plasticity,

comparable to that seen after heavy plastic deformation [68,76]. The research findings

showed an increase in the plastic strain of Zr62Cu24Fe5Al9 BMG from 4.9% to 7.6% under

uniaxial compression as well as a reduction in micro-hardness. Furthermore, the results

demonstrated improved plasticity of Cu46Zr46Al7Gd1 BMG from 1.4% in the as-cast state

to 5.1% after 10 successive cryo-thermal cycles showing proliferation of shear bands

compared to the as-cast condition where macroscopic plasticity is limited by premature

failure on one or a few dominant shear bands [76].

In addition to the above mentioned post-processing treatments to enhance the toughness of

glassy alloys, various practical surface modification techniques have been developed to

effectively improve the plasticity of metallic glasses [77]. Premised upon literature

investigations, the surface treatments utilized for metallic glasses to mitigate the

localization of strain are in principle classified into three types of categories, including

mechanical treatment [78–85], local thermal (laser surface) treatment [86] and surface

coating technique [87–90]. The main approach using these treatments is to introduce

compressive residual stresses in the surface region of the BMG, fabricating an amorphous

matrix composite with crystalline precipitated second phases in the surface region and

introducing a surface coating as geometrical lateral confinement, respectively. These

strategies can be applied to alter the deformation mode of BMGs from inhomogeneous

(localized shear banding) to homogeneous deformation (multiple shear band formation) at

room temperature. Zhang et al. [91] showed a dramatic effect of shot peening on the

plasticity of rod-shaped as-cast Zr-based BMG samples in uniaxial compression. It was

demonstrated that the compressive plasticity improved from the average 6%, maximum

7%, in as-cast BMGs to average 11%, maximum 22%, in peened samples (Fig. 2.7-a). They

reported that introducing compressive residual stresses in the surface region of the peened

samples, together with the pre-existing shear bands in the surface (as a consequence of

peening process), result in the multiplication of shear bands (high population of shear bands

Background

23

shown in Fig. 2.7-c), facilitating plasticity, rather than a premature failure on a few

dominant shear bands observed in Fig. 2.7-b.

Fig. 2.7. (a) Stress-strain curves for two as-cast, one abraded, and two shot-peened BMGs. (b) and (c)

scanning electron micrographs of areas close to the fracture surface (shaded in the set to “a”) indicating shear-

banding in as-cast and shot-peened BMGs, respectively [91].

Another surface engineering technique, extensively applied to crystalline metallic

counterparts, is thermochemical treatment including carburizing, nitriding, oxidizing, and

nitrocarburizing [92]. The thermochemical processing method is the changing of the

metal’s surface composition deliberately to make a hardened case by the dissolution of

interstitial elements constituting hydrogen (H), boron (B), carbon (C), nitrogen (N) and

oxygen (O) in the surface region of metallic materials at elevated temperature aiming at

enhancing the material’s performance [92]. On the other hand, this type of surface-

treatment has not been systematically investigated on amorphous alloys. Due to the

existence of elemental constituents in the chemical composition BMGs, for example,

ZrCuAl-based BMGs like Zr and Al which are strong oxide formers (more negative values

for fG of their oxide formation [93]) and they possess high affinity for oxygen, there

should be a potential for dissolving oxygen in their surface region. This can be applied as

an effective means for surface hardening as well as introducing compressive residual stress

in the BMG surface leading to decelerate the surface crack nucleation. It is suggested that

if the ratio of the volume of oxide to that of metal, i.e. Pilling-Bedworth ratio ( PBR= oxide

metal

V

V

Oxidation behavior of metallic glasses

24

) is greater than unity, the growth stresses are compressive [94]. Hence, to arrive at a better

understanding of how oxygen dissolution in the highly-reactive (or highly-oxidizable)

BMGs influences the surface microstructure as well as the growth of the residual stresses

in these regions, it is crucially important to study the oxidation behavior of such these

defect-free materials under various oxidizing conditions.

2.3 Oxidation behavior of metallic glasses

Many metallic alloys are thermodynamically unstable in an oxidizing atmosphere and tend

to form metal oxides in air (see Ellingham-Richardson diagram [95] shown in Fig. 2.8).

Whether oxides do develop depends on the oxygen partial pressure in the oxidizing

atmosphere [96].

Fig. 2.8. Ellingham-Richardson diagram for some oxides [95].

In crystalline oxide scales, the growth of the oxide layer depends on the transport properties

of anions (O2-) and cations (Men+), which takes place along lattice defects in the oxide

crystal lattice. In crystalline metals oxygen atoms (or ions) can diffuse by a vacancy

mechanism or by an interstitial mechanism. Furthermore, high-diffusivity paths as grain

boundaries and dislocations contribute to the transport of the species [97]. On the other

hand, in homogenous glassy metals which lack crystalline defects, oxygen transport can be

described in terms of excess free volume in the atomic structure, owing to the high

solidification rates during fabrication process as well as the existence of alloying elements

Background

25

with high atomic radius mismatch [98]. A high concentration of free volume provides

passages that facilitate the inward diffusion of oxygen. It is anticipated that oxygen

transport in the BMG is much faster than transport of metal species to the surface, due to

the absence of grain boundaries. Hence, the formation of an oxide scale at the surface is

unlikely and an internal oxidation zone develops instead. Oxygen transport can be more

complicated when an MG is exposed to the oxygen-containing atmosphere at temperatures

close to, within, and above the SLR (or ∆Tx) region, where relaxation occurs, followed by

devitrification of the amorphous structure along with oxygen ingress. Here it is noted that

the metastable MG (non-equilibrium state) relaxes slowly on heating to temperatures up to

Tg. This unfavorable effect results in the reduction of free volume and induces severe

brittleness [99]. It has also been reported that other physical properties of an MG will be

affected by sub-Tg annealing, including viscosity, specific heat, electrical resistivity, etc

[100]. On continued annealing of an MG to temperatures near Tx, crystallization

transformation by nucleation and growth of the crystalline intermetallic compounds to

reach an equilibrium state happens [101]. Moreover, the presence of various alloying

constituents in the chemical composition of the BMGs with different oxygen affinities

results in the development of a single- or multi-oxidation zones depending on the oxidation

process parameters including time, temperature and oxygen partial pressure (pO2). So far,

several studies were performed to investigate various factors influencing the oxidation

behavior of amorphous alloys containing at least two main metallic elements and additional

alloying components in various oxidizing environments. These studies are mostly applied

ex-situ (post mortem) and focus on the oxidation kinetics and the oxidation-induced

microstructure, i.e. thickness, surface morphology, and the composition of the oxide phases

formed in the surface region of the MGs [98,102].

Mondal et al. [103] investigated the oxidation behavior of the noble-metal containing

amorphous and nano-crystalline Zr70Pd30 and Zr80Pt20 ribbons using isothermal and non-

isothermal techniques. The results indicated that both nano-crystalline Zr-(Pd, Pt) melt-

spun are more resistant against oxidation compared with their amorphous state. This has

been attributed to the presence of a large degree of free volume in an amorphous structure

providing easy short-circuit paths for the diffusion of the oxygen. Another study has been

done to investigate the effect of atomic radius mismatch of alloying elements on the

oxidation resistance of binary amorphous Zr65M35 (M = Fe, Co, Ni, Cu, Rh, Pd or Au) alloys


26

at 573 K in static air [104]. It was found that an increase of the atomic size of the element

“M” increases the oxidation rate of the alloy, most probably due to the increase in the

atomic radius mismatch between Zr and the element “M” leading to the proliferation of the

available free space in the amorphous structure. As for instance, the amorphous Zr65M35

alloys containing Au (or Pd) with larger atomic sizes exhibit higher oxidation rates as

compared to the other investigated binary alloys. It was demonstrated that air-oxidation

resulted in the formation of (tetragonal and monoclinic) ZrO2 with finely distributed non-

oxidized Au (or Pd).

Zr53Ni23.5Al23.5 BMG (Tg ≈ 513°C) and crystalline alloys are exposed to the dry air and

pure oxygen experiment within the temperature range of 400°C-600°C to study the effect

of oxidation parameters on oxidation kinetics and microstructure evolution [105]. It has

been reported that the oxidation rates of the BMG are slightly higher than those of

crystallized BMGs in the temperature range of 400°C-500°C. The oxygen ingress to the

metallic glass substrate led to the development of an internal oxidation zone consisting of

t-ZrO2 and Al2O3 at 475°C while a minor amount of m-ZrO2 formed after oxidation at

500°C. An investigation conducted by Köster et al. [106] showed the surface response of a

ternary Cu-rich Cu60Zr30Ti10 BMG to an oxidizing atmosphere (synthetic air) at 320°C for

2 and 16 hr. They reported the formation of multi-oxidation zones in the surface of the

BMG which contain globular Cu-oxides decorated by the whisker-like features in the outer

oxidation zone and ZrO2 (and TiO2) lamellae embedding within the non-oxidized Cu in the

inner surface zone. It was claimed that the oxidation started with the formation of Cu2O

along the grooves of the ground surface followed by the formation of (Zr, Ti)O2. It has been

stated that the outer oxidation region formed because of outward diffusion of Cu from the

lamellar-structured inner oxidation zone to the surface. However, the driving force required

for the Cu transport to the surface remained undiscussed [106].

Moreover, various other research works were performed to study the microstructure

evolution during oxidation of the multicomponent BMGs (more than 3 main elements in

their chemical composition). The effect of the Zr content on the air-oxidation behavior of

Zr53Cu20Ni12Al10Ti5, Zr55Cu30Al10Ni5, and Zr65Cu15Al10Ni10 BMGs showing Tg, 384°C,

400°C, and 308°C at a heating rate of 0.167 Ks-1, respectively was investigated in the

temperature ranging from 300°C to 500°C [107]. It was observed that the oxidation kinetics

of the BMGs followed a linear rate law at 300°C and 350°C, suggesting surface reaction

Background

27

(gas-metal interface) was the rate determining step. However, at T>350°C, the oxidation

kinetics of the investigated BMGs followed a parabolic rate law, indicating an initial

relatively steep increase in the oxygen uptake, followed by a steady-state stage. Therefore,

the diffusion of oxygen is the rate-controlling step, implying that the oxidation rate

decreases with increasing the oxidation zone thickness. It was found that the oxygen

dissolution results in the formation of two oxidation zones mainly consisting of ZrO2 and

Cu-oxides in the internal and external surface zones, respectively. Partial devitrification

(formation of Zr-Cu and Zr-Al intermetallic compounds) of the BMG substrates caused by

the thermal oxidation experiment was also detected [107]. Mondal et al. [108] investigated

the effect of different initial microstructure (from amorphous to micro-crystalline) on non-

isothermal and isothermal air-oxidation of Zr58Cu28Al10Ti4 and Zr65Cu7.5Al7.5Ni10Pd10

BMGs. The results demonstrated that the oxidation resistance of Zr-based BMG having

different microstructure could be arranged in the following order: Crystalline

Zr>Crystalline Zr alloy>nano-quasicrystalline Zr alloy>relaxed amorphous Zr

alloy>amorphous Zr alloy. This has also been shown in the other studies where the

amorphous alloys exhibit poorer oxidation resistance compared with their crystalline

counterparts [109–111].

Although a significant number of studies was performed to clarify the general effect of

external (temperature, time and oxygen partial pressure) and internal (chemical

composition and microstructure) parameters on oxidation mechanism of MGs, the role of

residual stresses developed during oxidation on the growth of the surface-oxidation zones

has not been thoroughly justified for MGs. It is well established that oxidation is associated

with the development of residual stresses classified in (i) growth stresses generated by the

conversion of metal into oxide, generally leading to compressive residual stresses (ii)

thermal stresses developing during cooling due to the different thermal expansion

coefficients in oxide and metal [112,113]. Moreover, the long-term in-situ investigations

(direct-monitoring) on the development of the oxide phases in the surface zone of the

amorphous alloys have so far received no attention. Therefore, in order to address these

points regarding the underlying oxidation mechanism in MGs, more systematic

investigations are essentially required.


28

References

[1] E.D. Zanotto, J.C. Mauro, The glassy state of matter: Its definition and ultimate fate,

J. Non. Cryst. Solids. 471 (2017) 490–495.

[2] T. D. Bennett, J.C. Tan, Y. Yue, E. Baxter, C. Ducati, N.J. Terrill, H. H.M. Yeung,

Z. Zhou, W. Chen, S. Henke, A.K. Cheetham, G.N. Greaves, Hybrid glasses from

strong and fragile metal-organic framework liquids, Nat. Commun. 6 (2015) 1–7.

[3] G. Tarjus, Metal turned to glass, Nature. 448 (2007) 758–759.

[4] C. Suryanarayana, A. Inoue, Bulk Metallic Glasses, CRC Press, 2011.

[5] Q. Zheng, Y. Zhang, M. Montazerian, O. Gulbiten, J.C. Mauro, E.D. Zanotto, Y.

Yue, Understanding Glass through Differential Scanning Calorimetry, Chem. Rev.

119 (2019) 7848–7939.

[6] D. Turnbull, Under what conditions can a glass be formed?, Contemp. Phys. 10

(1969) 473–488.

[7] Z.P. Lu, Y. Li, S.C. Ng, Reduced glass transition temperature and glass forming

ability of bulk glass forming alloys, J. Non. Cryst. Solids. 270 (2000) 103–114.

[8] A. Inoue, Stabilization of metallic supercooled liquid and bulk amorphous alloys,

Acta Mater. 48 (2000) 279–306.

[9] A. Inoue, Chapter 14: Bulk amorphous alloys, Pergamon Mater. Ser. 2 (1999) 375–

415.

[10] A. Inoue, F.L. Kong, Q.K. Man, B.L. Shen, R.W. Li, F. Al-Marzouki, Development

and applications of Fe- and Co-based bulk glassy alloys and their prospects, J. Alloys

Compd. 615 (2015) S2–S8.

[11] H.S. Chen, Thermodynamic considerations on the formation and stability of metallic

glasses, Acta Metall. 22 (1974) 1505–1511.

[12] H.W. Kui, A.L. Greer, D. Turnbull, Formation of bulk metallic glass by fluxing,

Appl. Phys. Lett. 45 (1984) 615–616.

[13] A. Peker, W.L. Johnson, A highly processable metallic glass:

Zr41.2Ti13.8Cu12.5Ni10.0Be22.5, Appl. Phys. Lett. 63 (1993) 2342–2344.

[14] M. Telford, The case for bulk metallic glass, Mater. Today. 7 (2004) 36–43.

[15] D. Turnbull, Phase Changes, 1956.

[16] H.A. Davies, Metallic glass formation, Butterworth & Co (Publishers) Ltd., 1983.

[17] A. Takeuchi, A. Inoue, Classification of Bulk Metallic Glasses by Atomic Size


to Characterization of the Main Alloying Element, Mater. Trans. 46 (2005) 2817–

Background

29

2829.

[18] N.J. Grant, Scope and Trends of Developments in Rapid Solidification Technology.,

Elsevier B.V., 1985.

[19] H.H. Liebermann, Sample preparation: methods and process characterization,

Butterworth & Co (Publishers) Ltd., 1983.

[20] W.A. Heinemann, Rapid Solidification - Process Technology for New Materials and

Products., Elsevier B.V., 1985.

[21] B. Zheng, Y. Lin, Y. Zhou, E.J. Lavernia, Gas atomization of amorphous aluminum

powder: Part II. experimental investigation, Metall. Mater. Trans. B Process Metall.

Mater. Process. Sci. 40 (2009) 995–1004.

[22] B. V. Molotilov, T.N. Eremin, T.A. Shestakova, Two-Roller System for Producing

Fast-Quenched Metal Strip, Steel Transl. 48 (2018) 754–755.

[23] I.A. Bataev, S. Tanaka, I.D. Kuchumova, O.E. Matts, D.V. Lazurenko, K.

Hokamoto, A.A. Bataev, Metastable structures formed at the interface in explosively

welded materials, (2016).

[24] A. Inoue, N. Nishiyama, T. Matsuda, Preparation of bulk glassy Pd40Ni10Cu30P20

alloy of 40 mm in diameter by water quenching, Mater. Trans. Jim. 37 (1996) 181–

184.

[25] Q.S. Zhang, W. Zhang, A. Inoue, Ni-free Zr-Fe-Al-Cu bulk metallic glasses with

high glass-forming ability, Scr. Mater. 61 (2009) 241–244.

[26] G. Duan, A. Wiest, M.L. Lind, A. Kahl, W.L. Johnson, Lightweight Ti-based bulk

metallic glasses excluding late transition metals, Scr. Mater. 58 (2008) 465–468.

[27] M. Stoica, K. Hajlaoui, J. Das, J. Eckert, A.R. Yavari, FeNbB bulk metallic glass

with high boron content, Rev. Adv. Mater. Sci. 18 (2008) 61–65.

[28] A. Inoue, B. Shen, H. Koshiba, H. Kato, A.R. Yavari, Cobalt-based bulk glassy alloy

with ultrahigh strength and soft magnetic properties, Nat. Mater. 2 (2003) 661–663.

[29] P. Jia, H. Guo, Y. Li, J. Xu, E. Ma, A new Cu-Hf-Al ternary bulk metallic glass with

high glass forming ability and ductility, Scr. Mater. 54 (2006) 2165–2168.

[30] P. Ramasamy, A. Szabo, S. Borzel, J. Eckert, M. Stoica, A. Bárdos, High pressure

die casting of Fe-based metallic glass, Sci. Rep. 6 (2016).

[31] Y. Yokoyama, E. Mund, A. Inoue, L. Schultz, Production of Zr55Cu30Ni5Al10 glassy

alloy rod of 30 mm in diameter by a cap-cast technique, Mater. Trans. 48 (2007)

3190–3192.

[32] S. Cardinal, J.M. Pelletier, J.C. Qiao, G. Bonnefont, G. Xie, Influence of spark

plasma sintering parameters on the mechanical properties of Cu50Zr45Al5 bulk

metallic glass obtained using metallic glass powder, Mater. Sci. Eng. A. 677 (2016)


30

116–124.

[33] J. Schroers, Processing of bulk metallic glass, Adv. Mater. 22 (2010) 1566–1597.

[34] J. Schroers, T. Nguyen, S. O’Keeffe, A. Desai, Thermoplastic forming of bulk

metallic glass-Applications for MEMS and microstructure fabrication, Mater. Sci.

Eng. A. 448–451 (2007) 898–902.

[35] D. Herzog, V. Seyda, E. Wycisk, C. Emmelmann, Additive manufacturing of metals,

Acta Mater. J. 117 (2016) 371–392.

[36] S. Pauly, C. Schricker, S. Scudino, L. Deng, U. Kühn, Processing a glass-forming

Zr-based alloy by selective laser melting, Mater. Des. 135 (2017) 133–141.

[37] P. Bordeenithikasem, Y. Shen, H.L. Tsai, D.C. Hofmann, Enhanced mechanical

properties of additively manufactured bulk metallic glasses produced through laser

foil printing from continuous sheetmetal feedstock, Addit. Manuf. 19 (2018) 95–

103.

[38] H.Y. Jung, S.J. Choi, K.G. Prashanth, M. Stoica, S. Scudino, S. Yi, U. Kühn, D.H.

Kim, K.B. Kim, J. Eckert, Fabrication of Fe-based bulk metallic glass by selective

laser melting: A parameter study, Mater. Des. 86 (2015) 703–708.

[39] B. Sarac, J. Schroers, Designing tensile ductility in metallic glasses, Nat. Commun.

4 (2013) 1–7.

[40] C. Fan, R.T. Ott, T.C. Hufnagel, Metallic glass matrix composite with precipitated

ductile reinforcement, Appl. Phys. Lett. 81 (2002) 1020–1022.

[41] Y. Li, S.J. Poon, G.J. Shiflet, J. Xu, D.H. Kim, J.F. Löffler, Formation of bulk

metallic glasses and their composites, MRS Bull. 32 (2007) 624–628.

[42] S.F. Guo, L. Liu, N. Li, Y. Li, Fe-based bulk metallic glass matrix composite with

large plasticity, Scr. Mater. 62 (2010) 329–332.

[43] J. Qiao, H. Jia, P.K. Liaw, Metallic glass matrix composites, Mater. Sci. Eng. R

Reports. 100 (2016) 1–69.

[44] Z. Zhu, H. Zhang, Z. Hu, W. Zhang, A. Inoue, Ta-particulate reinforced Zr-based

bulk metallic glass matrix composite with tensile plasticity, Scr. Mater. 62 (2010)

278–281.

[45] B.Z. Ding, H.M. Fu, H.F. Zhang, Z.Q. Hu, H. Li, A.M. Wang, Synthesis and

characteristics of 80 vol.% tungsten (W) fibre/Zr based metallic glass composite,

Intermetallics. 17 (2009) 1070–1077.

[46] A.L. Greer, Metallic Glasses, Fifth Edit, Elsevier B.V., 2014.

[47] M.F. Ashby, A.L. Greer, Metallic glasses as structural materials, Scr. Mater. 54

(2006) 321–326.

Background

31

[48] M.D. Demetriou, M.E. Launey, G. Garrett, J.P. Schramm, D.C. Hofmann, W.L.

Johnson, R.O. Ritchie, A damage-tolerant glass, Nat. Mater. 10 (2011) 123–128.

[49] J. Schroers, W.L. Johnson, Ductile bulk metallic glass, Phys. Rev. Lett. 93 (2004)

20–23.

[50] D.C. Hofmann, Bulk Metallic Glasses and Their Composites: A Brief History of

Diverging Fields, J. Mater. 2013 (2013) e517904.

[51] A. Gebert, Corrosion behaviour of magnesium (Mg)-based bulk metallic glasses,

Corros. Magnes. Alloy. (2011) 207–233.

[52] J.R. Scully, A. Gebert, J.H. Payer, Corrosion and related mechanical properties of

bulk metallic glasses, J. Mater. Res. 22 (2007) 302–313.

[53] A.I. Salimon, M.F. Ashby, Y. Bréchet, A.L. Greer, Bulk metallic glasses: What are

they good for?, Mater. Sci. Eng. A. 375–377 (2004) 385–388.

[54] H.F. Li, Y.F. Zheng, Recent advances in bulk metallic glasses for biomedical

applications, Acta Biomater. 36 (2016) 1–20.

[55] C.A. Schuh, T.C. Hufnagel, U. Ramamurty, Mechanical behavior of amorphous

alloys, Acta Mater. 55 (2007) 4067–4109.

[56] M.K. Miller, P. Liaw, Bulk Metallic Glasses: An Overview, 2007.

[57] J.J. Lewandowski, M. Shazly, A.S. Nouri, Intrinsic and extrinsic toughening of

metallic glasses, Scr. Mater. 54 (2006) 337–341.

[58] S.J. Poon, A. Zhu, G.J. Shiflet, Poisson’s Ratio and Intrinsic Plasticity of Metallic

Glasses, Appl. Phys. Lett. 92 (2008) 2006–2009.

[59] S. V. Madge, D. V. Louzguine-Luzgin, J.J. Lewandowski, A.L. Greer, Toughness,

extrinsic effects and Poisson’s ratio of bulk metallic glasses, Acta Mater. 60 (2012)

4800–4809.

[60] P. Yu, H.Y. Bai, Poisson’s ratio and plasticity in CuZrAl bulk metallic glasses,

Mater. Sci. Eng. A. 485 (2008) 1–4.

[61] J.C. Lee, Y.C. Kim, J.P. Ahn, H.S. Kim, Enhanced plasticity in a bulk amorphous

matrix composite: Macroscopic and microscopic viewpoint studies, Acta Mater. 53

(2005) 129–139.

[62] J. Eckert, J. Das, K.B. Kim, F. Baier, M.B. Tang, W.H. Wang, Z.F. Zhang, High

strength ductile Cu-base metallic glass, Intermetallics. 14 (2006) 876–881.

[63] Y.K. Xu, H. Ma, J. Xu, E. Ma, Mg-based bulk metallic glass composites with

plasticity and gigapascal strength, Acta Mater. 53 (2005) 1857–1866.

[64] J. Eckert, J. Das, S. Pauly, C. Duhamel, Mechanical properties of bulk metallic

glasses and composites, J. Mater. Res. 22 (2007) 285–301.


32

[65] D.C. Hofmann, J.-Y. Suh, A. Wiest, G. Duan, M.-L. Lind, M.D. Demetriou, W.L.

Johnson, Designing metallic glass matrix composites with high toughness and

tensile ductility, Nature. 451 (2008) 1085–1089.

[66] K.W. Park, C.M. Lee, M. Wakeda, Y. Shibutani, E. Fleury, J.C. Lee, Homogeneous

deformation of bulk amorphous alloys during elastostatic compression and its

packing density dependence, Scr. Mater. 59 (2008) 710–713.

[67] T. Egami, K. Maeda, V. Vitek, Structural defects in amorphous solids A computer

simulation study, Philos. Mag. A Phys. Condens. Matter, Struct. Defects Mech.

Prop. 41 (1980) 883–901.

[68] A.L. Greer, Y.H. Sun, Stored energy in metallic glasses due to strains within the

elastic limit, Philos. Mag. 96 (2016) 1643–1663.

[69] Y. Sun, A. Concustell, A.L. Greer, Thermomechanical processing of metallic

glasses: extending the range of the glassy state, Nat. Rev. Mater. 1 (2016).

[70] D.J. Magagnosc, G. Kumar, J. Schroers, P. Felfer, J.M. Cairney, D.S. Gianola, Effect

of ion irradiation on tensile ductility, strength and fictive temperature in metallic

glass nanowires, Acta Mater. 74 (2014) 165–182.

[71] Y. Tong, T. Iwashita, W. Dmowski, H. Bei, Y. Yokoyama, T. Egami, Structural

rejuvenation in bulk metallic glasses, Acta Mater. 86 (2015) 240–246.

[72] H. Zhou, R. Hubek, M. Peterlechner, G. Wilde, Two-stage rejuvenation and the

correlation between rejuvenation behavior and the boson heat capacity peak of a bulk

metallic glass, Acta Mater. 179 (2019) 308–316.

[73] R. Raghavan, K. Boopathy, R. Ghisleni, M.A. Pouchon, U. Ramamurty, J. Michler,

Ion irradiation enhances the mechanical performance of metallic glasses, Scr. Mater.

62 (2010) 462–465.

[74] Y. Yokoyama, K. Yamano, K. Fukaura, H. Sunada, A. Inoue, Ductility improvement

of Zr55Cu30Al10Ni5 bulk amorphous alloy, Scr. Mater. Mater. 44 (2001) 1529–1533.

[75] S.C. Lee, C.M. Lee, J.W. Yang, J.C. Lee, Microstructural evolution of an

elastostatically compressed amorphous alloy and its influence on the mechanical

properties, Scr. Mater. 58 (2008) 591–594.

[76] S. V. Ketov, Y.H. Sun, S. Nachum, Z. Lu, A. Checchi, A.R. Beraldin, H.Y. Bai,

W.H. Wang, D. V. Louzguine-Luzgin, M.A. Carpenter, A.L. Greer, Rejuvenation of

metallic glasses by non-affine thermal strain, Nature. 524 (2015) 200–203.

[77] T.G. Nieh, Y. Yang, J. Lu, C.T. Liu, Effect of surface modifications on shear

banding and plasticity in metallic glasses: An overview, Prog. Nat. Sci. Mater. Int.

22 (2012) 355–363.

[78] F.O. Méar, B. Doisneau, A.R. Yavari, A.L. Greer, Structural effects of shot-peening

in bulk metallic glasses, 483 (2009) 256–259.

Background

33

[79] F.O. Méar, G. Vaughan, A.R. Yavari, A.L. Greer, G. Vaughan, A.R. Yavari, A.L.G.

Residual-stress, Residual-stress distribution in shot-peened metallic-glass plate,

0839 (2017).

[80] J. Fu, Y. Zhu, C. Zheng, R. Liu, Z. Ji, Optics & Laser Technology Enhanced

plasticity of bulk metallic glass in different aspect ratios via laser shock peening with

multiple impacts, Opt. Laser Technol. 83 (2016) 43–48.

[81] J. Fu, Y. Zhu, C. Zheng, R. Liu, Z. Ji, Applied Surface Science Effect of laser shock

peening on mechanical properties of Zr-based bulk metallic glass, Appl. Surf. Sci.

313 (2014) 692–697.

[82] A. Amanov, Y. Pyun, S. Sasaki, Tribology International Effects of ultrasonic

nanocrystalline surface modification (UNSM) technique on the tribological behavior

of sintered Cu-based alloys, Tribiology Int. 72 (2014) 187–197.

[83] C. Ma, H. Qin, Z. Ren, S.C.O. Keeffe, J. Stevick, G.L. Doll, Y. Dong, C. Ye,

Increasing fracture strength in bulk metallic glasses using ultrasonic nanocrystal

surface modification, 718 (2017) 246–253.

[84] Y. Cao, X. Xie, J. Antonaglia, B. Winiarski, G. Wang, Y.C. Shin, P.J. Withers, K.A.

Dahmen, P.K. Liaw, Laser Shock Peening on Zr-based Bulk Metallic Glass and Its

Effect on Plasticity : Experiment and Modeling, Nat. Publ. Gr. (2015) 1–8.

[85] J. Fan, A. Chen, J. Wang, J. Shen, J. Lu, Improved plasticity and fracture toughness

in metallic glasses via surface crystallization, Intermetallics. 19 (2011) 1420–1427.

[86] G. Wu, R. Li, Z. Liu, B. Chen, Y. Li, Y. Cai, T. Zhang, Induced multiple

heterogeneities and related plastic improvement by laser surface treatment in CuZr-

based bulk metallic glass, Intermetallics. 24 (2012) 50–55.

[87] W. Chen, K.C. Chan, P. Yu, G. Wang, Encapsulated Zr-based bulk metallic glass

with large plasticity, Mater. Sci. Eng. A. 528 (2011) 2988–2994.

[88] M. Meng, Z. Gao, L. Ren, H. Yang, S. Ma, Materials Science & Engineering A

Improved plasticity of bulk metallic glasses by electrodeposition, Mater. Sci. Eng.

A. 615 (2014) 240–246.

[89] S.B. Qiu, K.F. Yao, Novel application of the electrodeposition on bulk metallic

glasses, Appl. Surf. Sci. 255 (2008) 3454–3458.

[90] W. Chen, K.C. Chan, S.H. Chen, S.F. Guo, W.H. Li, G. Wang, Plasticity

enhancement of a Zr-based bulk metallic glass by an electroplated Cu / Ni bilayered

coating, 552 (2012) 199–203.

[91] Y. Zhang, W.H. Wang, A.L. Greer, Making metallic glasses plastic by control of

residual stress, 5 (2006) 857–860.

[92] E.J. Mittemeijer, M.A.J. Somers, Thermochemical Surface Engineering of Steels,

Woodhead Publishing Limited, 2015.


34

[93] Ihsan Barin, Thermochemical Data of Pure Substances, Wiley-VCH Verlag, 1995.

[94] N. B. Pilling and R. E. Bedworth, The oxidation of metals at high temperatures, J.

Inst. Met. XXIX (1923) 529–582.

[95] M. Hasegawa, Ellingham Diagram, Elsevier Ltd., 2013.

[96] F.S.P. N. Birks, G.H. Meier, Introduction to the high-temperature oxidation of

metals, 2nd Editio, Cambridge University Press, 2006.

[97] A. Atkinson, Wagner theory and short circuit diffusion, (1988) 1045–1051, Vol. 4.

[98] Z. Xu, Y. Xu, A. Zhang, J. Wang, Z. Wang, Oxidation of amorphous alloys, J. Mater.

Sci. Technol. 34 (2018) 1977–2005.

[99] P. Murali, U. Ramamurty, Embrittlement of a bulk metallic glass due to sub-Tg

annealing, Acta Mater. 53 (2005) 1467–1478.

[100] T. Egami, Structural Relaxation in Metallic Glasses, Ann. N. Y. Acad. Sci. 371

(1981) 238–251.

[101] S. Haratian, M. Haddad-Sabzevar, Thermal stability and non-isothermal

crystallization kinetics of Ti41.5Cu42.5Ni7.5Zr2.5Hf5Si1 bulk metallic glass, J. Non.

Cryst. Solids. 429 (2015) 164–170.

[102] M. Zhang, L. Deng, D. Yao, J. Jin, X. Wang, Multilayered scale formation during

Zr-based metallic glass oxidation in the supercooled liquid region, Corros. Sci. 111

(2016) 556–567.

[103] K. Mondal, U.K. Chatterjee, B.S. Murty, Oxidation behavior of amorphous and

nanoquasicrystalline Zr-Pd and Zr-Pt alloys, J. Alloys Compd. 460 (2008) 172–181.

[104] H.M. Kimura, K. Asami, A. Inoue, T. Masumoto, The oxidation of amorphous Zr-

base binary alloys in air, 35 (2000) 909–915.

[105] W. Kai, H.H. Hsieh, Y.R. Chen, Y.F. Wang, C. Dong, Oxidation behavior of an

Zr53Ni23.5Al23.5bulk metallic glass at 400-600 °C, Intermetallics. 15 (2007) 1459–

1465.

[106] U. Köster, L. Jastrow, M. Meuris, Oxidation of Cu60Zr30Ti10 metallic glasses, Mater.

Sci. Eng. A. 448–451 (2007) 165–168.

[107] H.H. Hsieh, W. Kai, R.T. Huang, M.X. Pan, T.G. Nieh, Effect of Zr-content on the

oxidation and phase transformation of Zr-base amorphous alloys in air,

Intermetallics. 12 (2004) 1089–1096.

[108] K. Mondal, U.K. Chatterjee, B.S. Murty, Oxidation behavior of multicomponent Zr-

based amorphous alloys, J. Alloys Compd. 433 (2007) 162–170.

[109] W. Kai, T.H. Ho, H.H. Hsieh, Y.R. Chen, D.C. Qiao, F. Jiang, G. Fan, P.K. Liaw,

Oxidation behavior of CuZr-based glassy alloys at 400 °C to 500 °C in dry air,

Background

35

Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 39 (2008) 1838–1846.

[110] W. Kai, P.C. Lin, W.S. Chen, C.P. Chuang, P.K. Liaw, H.H. Huang, H.H. Hsieh,

Air-oxidation of a Zr50Cu43Al7 bulk metallic glass at 400-500°C, Corros. Sci. 64

(2012) 98–104.

[111] Y. Wu, T. Nagase, Y. Umakoshi, Effect of crystallization behavior on the oxidation

resistance of a Zr-Al-Cu metallic glass below the crystallization temperature, J. Non.

Cryst. Solids. 352 (2006) 3015–3026.

[112] J. Stringer, Stress generation and relief in growingoxide films, Corros. Sci. 10 (1970)

513–543.

[113] D.L. Douglass, The role of oxide plasticity on the oxidation behavior of metals: A

review, Oxid. Met. 1 (1969) 127–142.


36

Materials and experimental methods

37

3 Materials and experimental methods

In this chapter, an overview of the materials and experimental methods employed in this

project is provided. The methods, which were utilized, can be classified in techniques for

materials microstructure characterization and methods to investigate the development of

the macroscopic strain (and residual stresses) in the surface region of the oxidized ZrCuAl-

based BMG samples. For the characterization, X-ray diffraction (XRD) and advanced

(electron) microscopy techniques have been used. In order to monitor the microstructural

changes, phase composition, and residual stress evolution during air-oxidation of the

BMGs, in-situ XRD was applied. The evolution of macroscopic strain (and residual

stresses) in the thermochemically surface-engineered BMG was evaluated from the ex-situ

lattice strain X-ray diffraction sin2ψ and incremental ring-core focused ion beam milling

(FIB) methods. A general introduction to each technique applied during this project is given

in the following sections. Detailed descriptions of the applied parameters can be found in

the results chapters.

3.1 Materials and gaseous oxidizing treatment

A series of ZrCuAl-based bulk amorphous alloys with nominal chemical composition

Zr51.3Al8.5Cu31.3Ni4Ti4.9, (Zr55Al10Cu30Ni5)98Er2, and Zr48Cu36Ag8Al8 (at. %) were

investigated in this research work. These BMGs are known as good glass-forming alloys

with a potential engineering application. At the same time, the presence of Zr and Al in

their alloying composition, which both have a high affinity for oxygen, enables the

dissolution of an appreciable amount of oxygen in the surface region of the investigated

BMGs. All the BMGs were provided by the Jiangxi Academy of sciences in China and

were fabricated using vacuum arc-melting of high purity (99.9 wt. %) elemental

constituents in a Ti-gettered argon atmosphere. The synthesized ingots were re-melted

several times to ensure compositional homogeneity. The homogenous alloying ingots were

subsequently cast into a plate with dimensions ~2×10×60 mm3 using the copper mold

casting method (Rapid Quench Machine System VF-RQT50, Makabe Co. Ltd. Japan). The

amorphous structure of the investigated materials was initially examined using X-ray

diffractometry (Bruker D8 discover) with CuKα radiation with a characteristic wavelength

of λ=0.15418 nm and recorded in the 2θ range 25°-90°. The X-ray diffractograms of the

as-cast ZrCuAl-based BMGs are presented in Fig. 3.1. The presence of two amorphous

Materials and gaseous oxidizing treatment

38

humps appearing at 2θ≈38° and 65° for all the investigated BMGs confirms the glassy state

without detectable crystallinity.

Fig. 3.1. XRD diffractograms of the investigated as-cast ZrCuAl-based BMGs.

The thermal stability and crystallization behavior of the mentioned bulk glassy alloys were

determined by continuous heating at a heating rate of 10 K/min in a Netzsch STA 449C

differential thermal analyzer (DTA) under a flow of purified argon at a flow rate of 50

cm3/min. About 40 mg of the glassy samples was cut from the as-cast BMGs for each

measurement. An alumina (Al2O3) crucible was utilized for continuous heating in the DTA

apparatus. In the DTA method, the temperature difference between a specimen (a few mg

of the glassy sample) and a reference sample subjected to the specific thermal programs

with a controlled heat flow is measured using a differential thermocouple1. The method

enables monitoring the heat evolved (exothermic) or absorbed (endothermic) when an

investigated material transforms. The variation of heat change is plotted on the Y-axis and

the temperature (or time) on the X-axis during isochronal (or isothermal) heating [1].

1 The differential thermocouple is in contact with the underside of the specimen and

reference crucibles [1].


39

The as-cast BMG samples were oxidized using different oxidizing parameters (time and

temperature) and it performed in a proprietary gas mixture providing either an extremely

low or an extremely high oxygen partial pressure (pO2). The focus was on applying

oxidizing at temperatures below Tg in order to prevent devitrification of the BMGs. The

BMGs were exposed to gaseous oxidizing, which consisted of heating the 3×3×1 mm2

specimen non-isothermally at a heating rate of 20 K/min to a specific temperature followed

by isothermal holding for some duration and finally cooling to ambient temperature at a

cooling rate of 15 K/min. This has been carried out in a thermo-gravimetric analyzer (TGA-

Netzsch STA 449C) which enabled to continuously measure the oxygen uptake (or mass

gain) of a BMG specimen monitored by a thermobalance as a function of oxidizing

duration. The TGA technique has a resolution of 1 μg and is useful for those

transformations involving the absorption or evolution of gases from a specimen and

configured for vacuum or variable atmospheres [1].

In addition to the oxidizing treatment in the controlled atmosphere, the long term oxidation

of the BMGs was followed in-situ and ex-situ using X-ray diffraction (Bruker D8) and

thermogravimetry (Netzsch STA 449C) in an air atmospheric condition, respectively. This

has been conducted in order to understand the governing oxidation mechanism and the

crystalline oxide phases developed in the surface of the multi-component ZrCuAl-based

BMGs. Prior to oxidation, the BMG plates were sliced into rectangle-shape samples using

a Struers Accutom-50. Irrespective of the treatment, the surface of all specimens was

manually polished to a final step of 1 µm diamond using Struers DP-U2 and cleaned with

ethanol in an ultrasonic bath. A detailed description of oxidizing parameters and sample

preparation is given in the experimental sections of each result chapter.

3.2 Characterization

X-ray diffraction

X-ray diffraction is among the non-destructive highly-used methods for the identification

and characterization of the crystalline materials by measuring the interplanar lattice

spacing. The acquired X-ray diffractogram demonstrates a characteristic of the periodic

atomic arrangement of a material and can be used for determining its crystal structure. In a

diffractometer setup, the X-rays [2] are produced by heating a tungsten filament leading to

the emission of electrons. These electrons are then accelerated by applying a voltage

Characterization

40

towards an anode-target material. When electrons have enough energy to remove inner

shell electrons of the anode source, characteristic X-rays are created [3,4].

The interaction of X-rays with the crystalline matter yields constructive interference when

Bragg’s condition is fulfilled [4,5]:

2 sinhkln d 3-1

where n is an integer, λ is the wavelength of the X-rays, dhkl is the lattice spacing in a

crystalline material causing the diffraction for the hkl lattice planes, and θ is the incident

beam angle. The Bragg equation correlates the wavelength of X-ray radiation to the

diffraction angle and the interplanar lattice spacing. Diffraction takes place by constructive

interference of a beam of X-rays scattered at particular angles from each set of

crystallographic planes [4], which is schematically presented in Fig. 3.2. Here it is noted

that for an amorphous material lacking the long-range arrangement of atoms in a 3D

structure, the X-rays are scattered in many directions giving rise to the presence of a large

hump (a very broad peak) located in a wide scattering range compared with high-intensity

narrow diffraction peaks obtained for crystals in their diffractograms (cf. Fig. 3-1).

Fig. 3.2. Schematic illustration of Bragg diffraction by a set of crystallographic planes. The incident X-ray

beams with wavelength, λ are constructively scattered by crystallographic planes (n=0, 1, 2, …) having

interplanar lattice spacing, dhkl. This shows that the values of λ, dhkl, and θ satisfy Bragg’s law condition. The

diffraction vector, g is perpendicular to the crystallographic planes. The figure is reprinted and recreated from

[6].


41

In this technique, the sample is scanned in a range of 2θ (scattering angle), where all

possible diffraction directions can be obtained. The diffracted X-ray signal is recorded by

a detector and then converted to an acquired count rate (intensity), which is output to a

device [4]. The geometry of an X-ray diffractometer used for the ex-situ phase

identification in this study is such that the stage moves in the path of the X-ray beam at an

angle θ while the detector is attached on an arm to collect the diffracted X-rays and moves

at an angle of 2θ. The goniometer is utilized to keep the angle and movement of the stage.

The qualitative phase identification is generally obtained by comparing the interplanar

lattice spacing (or diffraction angle) with standard reference patterns [3,4].

It should be pointed out that the penetration depth of the applied X-rays into the material is

limited and it is within the range of a few micrometers. The lattice spacing determined by

XRD is diffracted-intensity weighted over a depth range, so-called information depth, τ,

from which 63% of the diffracted X-ray beam originates. In a symmetrical X-ray beam path

geometry, τ depends strongly on the diffraction angle, θ and the tilt angle, ψ (angle between

the lattice plane normal and diffraction vector) and can be obtained from [7,8]:

sin cos

2

3-2

where μ is the linear absorption coefficient of the analyzed sample composition for the

applied X-ray wavelength. In this study, ex-situ (post-mortem) angular dispersive X-ray

diffraction was extensively utilized in order to qualitatively characterize the phases formed

after gaseous oxidizing of the BMG samples. All measurements were performed using

Bruker D8 X-ray diffractometer equipped with a Cu-anode in parallel beam configuration,

operating at a voltage of 40 kV and a current of 40 mA. The diffractograms were recorded

in the 2θ range 25˚-90˚ and the step size and counting time per step were 0.04° (∆2 and

8 s, respectively. The crystalline phases formed as a result of oxidizing of the BMG samples

were identified by comparing the diffraction peak positions with the “Joint Committee on

Powder Diffraction Standard (JCPDS)” database (PDF Main Ex library Version 6.0.100.)

using the commercially available Bruker D8 EVA software (Version 6, 0, 0, 1- Copyright

SOCABIM 1996-2000).

Moreover, X-ray diffraction was applied using a conventional laboratory X-ray

diffractometer equipped with Cr radiation aiming at directly monitoring (in-situ) the

evolution of various crystalline oxide phases in the surface-adjacent region of the materials.

Characterization

42

Fig. 3.3 indicates the laboratory X-ray diffraction setup (θ-θ configuration) used for in-situ

air-oxidation experiments2. The in-situ XRD oxidation experiment3 was performed to a

BMG sample of dimensions 3×1×0.5 mm3. For this purpose, the sample was mounted on a

PtRh-heating band using two Alumina (Al2O3) rods. A heating controller connection was

attached to the heating stage and a thermocouple was mounted on the sample in order to

screen the temperature during air-oxidation. The thermal cycle included heating the sample

up to 600 K in temperature steps of 20 K and a diffractogram was acquired in each step.

The heating rate in-between the steps during isochronal heating was set to 60 K/min. After

reaching 600 K, the oxidation was followed isothermally for 60 hr. Cooling to room

temperature occurred in 50 K intervals. The cooling rate in-between the steps was 60

K/min. The X-ray acquisitions were done in the 2θ-range 42°-75°. This angular range was

chosen to cover all possible crystalline oxide phases, which may form during air-oxidation

of the investigated BMGs. The step size was 0.06° (∆2 and the counting time per step

was 3 s.

Fig. 3.3. Experimental setup for in-situ XRD air-oxidation representing how the 3×1×0.5 mm3 BMG sample

was mounted on a PtRh-heating band. A thermocouple was positioned on the sample to monitor the BMG

sample surface temperature.

2 This setup has also been utilized for in-situ lattice strain X-ray diffraction sin2ψ analysis.

A detailed information of the residual stress measurement parameters can be found in the

section 3.3.1 of this chapter.

3 The temperature accuracy during in-situ XRD air-oxidation experiment was ±10 K.


43

Microscopical characterization

Since the in-situ (and ex-situ) laboratory XRD method could not give information at a

sufficient depth to evaluate the microstructure evolution deeper into the material, several

(electron) microscopical characterization techniques; reflected light optical-, scanning- and

transmission electron microscopy (LOM, TEM and SEM) were applied. Such information

was obtained ex-situ (post-mortem), which revealed the distribution of various crystalline

phases, both oxidic and metallic phases. This section will give a general principle and an

overview of the sample preparation, electron transparent TEM specimen preparation, the

microscopy techniques, and the used experimental instruments in the present Ph.D. work.

Reflected light optical microscope (LOM)

A reflected light optical microscope, i.e. widely applied to investigate the microstructure

of metallic samples creates a magnified image of a sample (object) using the objective lens

and magnifies the image more with an eyepiece allowing the user to visualize it by the

naked eye [9]. In this technique, reflected light from the object first converges at the

objective lens and is focused at the position less than the focal length of the eyepiece

(second lens) to form a magnified primary inverted image. The light rays are further

converged by the eyepiece to make a final magnified virtual image on the human eye retina,

as schematically depicted in Fig. 3.4 [10].

In the present work, LOM was applied to acquire the oxidized sample’s cross-sectional

images under relatively low magnifications. To investigate the cross-section of the oxidized

BMGs, the samples were embedded using cold mounting resin. The mounted oxidized

BMGs were automatically ground and polished to 1 µm diamond using Struers RotoPol-

22. Bright-field (BF) LOM micrographs were collected using a Neophot 32 (Zeiss, Jena)

microscope where the perpendicular light reflection to the sample was viewed.

Characterization

44

Fig. 3.4. Schematic illustration of the path of light in an optical microscope with an eyepiece leading to the

visualization of the final image by a human eye retina. Reprinted and recreated from [10].

Micro-indentation

The hardness of a material is described as its resistance against plastic deformation under

indentation and several standard techniques are available to evaluate it. The main

differences of such techniques are the geometry of the indenter and the amount of the

applied force [11]. In the case of this work, the Vickers hardness (HV) test was used. In

this method, the HV value is obtained by dividing the applied force, P by the surface area

of the pyramidal depression which yields [11]:

2

2HV sin

2HV

P

d

3-3

where dHV is the average diagonal length of an indent left on the sample surface using a

diamond pyramid indenter with a face angle of α=136°. According to Eq. 3-3, the HV value

is independent of the magnitude of the applied force [11]. It is however noted that the HV

value generally increases with reducing the load. This is attributed to the surface

deformation and relaxation.

In this project, the Vickers micro-hardness profiles (standard E 384-17 [12]) of the oxidized

case were assessed using the Future-Tech FM700 instrument. The measurements were

carried out using a Vickers indenter with a 5-gf load and a dwell time of 10 s.


45

Scanning electron microscopy (SEM)

The scanning electron microscope is one of the most widely used instruments for the

material’s microstructure characterization. The interaction of the accelerated electrons and

the surface of the material establishes the basis for imaging and qualitative (and semi-

quantitative) chemical analysis in SEM [13,14]. As high energy electrons interact with the

atoms at the surface of the matter, various physical interactions occur which are

schematically depicted in Fig. 3.5. The volume and depth of the interaction region increases

with an increase of the electron beam energy or a decrease in an atomic number of the

elements in the specimen. The electron interaction results in elastic and inelastic scattering

events [10,15]. Elastic scattering is characterized by the insignificant energy loss as a result

of the deflection of the incident negatively charged electron by the positively charged atom

nucleus of the matter. Some of these electrons can leave the sample again and can be

detected as backscattered electrons (BSE) with almost the same energy as the incident

beam. Inelastic scattering generates secondary electrons (SEs) having low-energy (2-50

eV). Both SEs and BSEs signals generated by the electron scattering are used for making

SEM images. SEs are useful for obtaining detailed topographic contrast since they can only

escape from about 5-50 nm below the surface (cf. Fig. 3.5). However, BSEs are

advantageous for obtaining the elemental composition contrast and can escape from a depth

of 50-300 nm, depending on the energy of the primary electrons and the atomic number

and density of the sample (Fig. 3.5). The number of BSEs, which can escape from the

sample surface defined as the backscatter coefficient (ηBSE) depends on the atomic number

of the matter. Consequently, regions with a higher (average) atomic number back-scatter

more electrons and appear bright as compared to regions with low (average) atomic number

[15].

Here it is noted that the image resolution4 is controlled by the volume from which electrons

contribute to the image formation. This implies that the spatial resolution of SE images is

4 The smallest distance between two points that can be distinguished. The resolution

achieved by the electron microscopes is greater than that obtained for optical microscopes

where the light ray having longer wavelength compared with electrons is used for

Characterization

46

close to the spot size and is also considerably better than BSE images. Essentially, choosing

a higher accelerating voltage provides better resolution. On the other hand, it leads to an

increase in the interaction volume, thus electrons from deeper regions contribute to the

image formation. This can obscure the surface microstructural details and results in more

edge effects. The other point which needs to be addressed is that the final image resolution

can also be influenced by the aperture size and working distance5. It is known that small

aperture size and long working distance is beneficial for increasing the depth of field, which

in contrast leads to the negative impact on image resolution. Therefore an optimized

combination of accelerating voltage, aperture size, and working distance (preferably

intermediate aperture size and working distance) could be considered for achieving a good

combination of image resolution and depth of field [10,16].

Generally, SEM imaging is accompanied with energy-dispersive X-ray spectroscopy

(EDS) for chemical analysis. This enables the identification of radiation of specific energy

(or wavelength) for elemental analysis of the sample. The high energy particle (X-ray

photon, electron, or neutron) collides an electron in the inner shell of an atom and thus it

could dislodge an electron from its original energy level. The kicked-out electron leaves

the atom as a “free” electron and the atom becomes excited (unstable state). Since the atom

is in an ionized state, it tends to return to its normal state by filling the inner electron

vacancy with an electron that has lower binding energy in the atom. The energy difference

between an outer and an inner shell electron (surplus released energy) generates

characteristic X-ray radiation (cf. Fig. 3.5). The energy of characteristic X-rays is known

and relies on the atomic number of the element and used for elemental (chemical) analysis

[10,17].

illumination. The resolution which can be obtained by SEM is in order of 1-20 nm, while

resolution of LOM is typically around 200 nm.

5 The aperture is mainly used for limiting the divergence of the electron beam in its optical

path. In addition, the distance between the aperture and the sample surface is defined as

working distance [10].


47

Fig. 3.5. The interaction (excitation) volume of electrons and sample atoms below its surface showing the

origins of different signals and is highly influenced by the accelerating voltage and atomic number.

In this study, SEM was predominantly applied in combination with EDS6 for imaging from

the top surface morphology of the oxidized BMGs (SE mode), cross-sectional observation

(BSE mode), and elemental analysis. For this purpose, several SEM microscopes fitted with

EDS detectors were used in this work in which their model and the type of electron source7

are listed in table 3.1. For the top surface observation, the oxidized BMG samples were

mounted on SEM pin stubs using silver paint. For the cross-sectional investigations, the

cold-embedded samples were initially carbon sputter-coated using a Q150R E carbon

coater with carbon fiber (model code: E443) as a source. This has been conducted to ensure

the electron charge transfer on the dielectric epoxy.

Table 3.1. List of scanning electron microscopes (SEM) used in this Ph.D. project.

SEM model Filament EDS detector Location

JEOL JSM-5900 LaB6 Oxford (Inca X-act) MTU, DTU

Merlin, Carl Zeiss FEG-SEM Bruker Xflash 6ǀ60 Risø, DTU

Supra 35, Zeiss FEG-SEM Thermo NORAN system 6 MTU, DTU

FEI Helios Nanolab 600* FEG-SEM EDAX SD Apollo 10 Pegasus System CEN, DTU

FEI Helios Nanolab G3 CX* FEG-SEM EDAX SD Apollo 10 Pegasus System EMC, UOW * These two microscopes are dual-beam focused Ga+ beam-scanning electron microscope systems (FIB-

SEM) used for TEM sample preparation and incremental ring-core milling for the assessment of the surface

strain relaxation (and micron-scale residual stress determination), respectively.

6 The instrument was operated at 15-20 kV accelerating voltage for EDS acquisition on

oxidized ZrCuAl-based BMGs.

7 An enhanced electron brightness, about 100× greater than that obtained from a typical

tungsten (and LaB6) electron source could be achieved by field emission gun (FEG) source.

This can lead to the formation of electron probe size smaller than 2 nm, which provides

much higher resolution for SEM image [14].

Characterization

48

Transmission electron microscopy (TEM)

Transmission electron microscope (TEM) is a powerful and versatile material

characterization technique as it allows detailed microstructural imaging and the

identification of the crystallography of the crystalline phases. In principle, TEM enables

generating images of material microstructures with much higher magnification and

resolution than light and scanning electron microscopes. In TEM, electrons with high

interaction power are used instead of (visible) light as a source of illumination and they are

transmitted through an electron transparent thin foil (or lamella). Electrons with an energy

of 200 kV, have a wavelength in the order of ~ 0.025 Å (the wavelength of an electron is

inversely proportional to the acceleration voltage). The high resolution of TEM (0.1-0.2

nm) results from the short wavelengths of the electrons used for the microscope

illumination system. Such high resolution makes TEM extremely useful for directly

revealing ultrafine details of material microstructure. TEM can be operated either to

produce a diffraction pattern (DP) from a specific region (diffraction mode) or

microstructural images (image mode), where a comprehensive crystallographic correlation

can be made between the diffraction pattern and the corresponding crystalline phase. In the

diffraction mode, a diffraction pattern (DP) is formed in the back focal plane of the

objective lens, while an image is created on the image plane of the objective lens. The

projection and intermediate lenses below the objective lens are used to focus and magnify

either the diffraction pattern or the image onto a fluorescent screen [10,17–19].

The TEM consists of a column divided into two main parts: (i) the illumination system

comprises of the electron gun and the condenser lenses which control the electron beam

diameter and convergence angle. (ii) the imaging system which projects the transmitted

electrons onto a fluorescent screen or charged-coupled device (CCD) detector. Between

these two parts, the specimen holder stage is placed. The specimen stage controls the

position and rotation of the sample. The objective lens forms the image transmitted through

the imaging system [10]

Electron diffraction

An electron passing through the TEM lamella can be diffracted by a crystallographic lattice

plane at specific angles to the incoming electrons when Bragg’s condition (nλ=2dhklsinθ) is

satisfied. This can only occur from those crystallographic planes which are almost parallel


49

to the incident beam because a short wavelength corresponds to low diffraction angles.

Electron diffraction is not only used for acquiring microstructure images due to the

diffraction contrast but also for analyzing crystal structure (phase analysis, dspacing, ...)

[17,18]. The type of selected area electron diffraction (SAED) patterns acquired from a

material depends on its structure, orientation, and crystallite size in the sample. Fig. 3.6-a

shows a SAED pattern of an amorphous material demonstrating a diffuse halo in its pattern

as a consequence of the scattering of electrons in many directions by a non-periodic atomic

structure. Fig. 3.6-b shows a spot pattern acquired from a single crystal (or grain) oriented

such that several sets of planes are parallel to the beam result in a diffraction pattern

consisting of a regular array of spots. Moreover, a continuous ring pattern (several discrete

spots) that consist of a large number of grains forming a polycrystalline sample can also

be obtained as given in Fig. 3.6-c. This implies that if the sample contains several crystals

(or grains) having different orientations, the resulting diffraction pattern is the sum of the

individual patterns. In other words, diffraction occurs at certain crystallographic planes

with specific lattice spacing, and the spots are not randomly distributed but rather obey the

ring pattern (each ring has constant dspacing) [20].

Fig. 3.6. SAED pattern obtained from (a) amorphous material (b) single crystal (or grain) sample (c)

polycrystalline sample showing diffuse halo, spot, and ring patterns, respectively [20].

Imaging

The phase and amplitude of the electrons change during the interaction with the specimen

resulting in the formation of different contrast indicating different electron intensity. Two

main contrasts are responsible for creating an image depending on the thickness, atomic

number, and the crystal structure of the sample:

- Mass-thickness contrast

Essentially, a fraction of the accelerated electrons is absorbed by the material, so that the

intensity of the transmitted beam is reduced with an increment in the sample thickness.

Characterization

50

Also, the electron intensity is reduced with a high absorption coefficient. The absorption

coefficient is a function of the atomic number and the radiation energy. In other words, the

electrons are scattered more in the thick and dense areas of the specimen, leading to the

decrease in the electron density transmitted through the TEM lamella. Therefore, a decrease

of the electron intensity in the TEM images (appearing as dark regions) is observed. The

mass-thickness contrast is widely used to investigate the non-crystalline materials [10,18].

- Diffraction contrast

Diffraction contrast is simply defined when constructive electron scattering takes place at

a specified angle known as the Bragg angle. The diffraction contrast creates TEM bright-

field (BF) and dark-field (DF) images. TEM BF imaging produces an image using a

transmitted electron beam while in DF mode, an aperture blocks transmitted electron, and

only diffracted electrons are collected. It is noted that the diffraction contrast is highly

sensitive to sample tilting under different zone axes8, while the mass-thickness contrast is

only affected by changing the atomic number and density [10,18].

In the current project, TEM investigations were mainly carried out by combining BF-

imaging, EDS measurements, and SAED 9 to obtain more detailed microstructural

information from the oxidized regions in the surface of the oxidized BMGs. For this

purpose, a JEOL 3000 equipped with a FEG electron source operated at 300 kV and an

Oxford Instruments EDS detector at DTU Risø campus was used. All electron transparent

TEM foils were extracted in the direction perpendicular to the surface of the oxidized

BMGs. For TEM lift out, an FEI Helios NanoLab 600 dual beam SEM equipped with a

Ga+ ion beam source at DTU CEN was utilized. Initial rough FIB milling was performed

with a 30 kV ion beam acceleration voltage and 20 nA ion current. The area of interest on

the sample surface was protected from the excessive ion damage with the Pt-deposited

layer. Subsequently, the lamella was lifted out with the Omniprobe and then mounted to a

TEM Cu-grid. An ion-beam current of 90 pA-0.9 nA was used for the final thinning of the

8 A crystallographic direction contained by at least two unique crystallographic planes.

9 It should be mentioned that the indexing of the SAED patterns was performed using the

JEMS software.


51

lamella. A low acceleration voltage (2 kV) cleaning was ultimately carried out with a 24

pA ion-beam current to remove the residual Ga+ ion damage on the TEM lamella.

3.3 Residual stress determination

Residual stresses (RS) are defined as internal stresses within a material and are independent

of externally applied forces. RS are generated within a component as a consequence of

thermal or elastic mismatch, inhomogeneous plastic strain, or phase transformation. These

stresses remain in a component after manufacturing or post-processing and can lead to

premature failure and degrade material’s performance in service exposure [21]. On the

other hand, they can be beneficial where for instance compressive residual stresses are

deliberately engineered in the surface region of materials to enhance their resistance against

crack initiation at the surface and premature failure during mechanical loading, thereby

making its service life longer. RS are characterized based on the length scale over which

they equilibrate and can be classified into macro- and micro stresses [8,22]. Macro-stresses

(type I) are homogeneous over the entire material volume, while micro-stresses vary locally

at the scale of individual grains (type II) or as a consequence of defects within the grains

(type III). Since the residual stresses can significantly influence the performance of

engineering materials, it is crucially important to their origin, magnitude, and distribution

in a component. For this purpose, several methods were developed to determine RS which

are classified into non-destructive and (semi-) destructive methods [8,22,23].

The exposure of ZrCuAl-based BMGs to an oxygen-carrying atmosphere at (moderately)

high temperatures would lead to the development of oxidized zones in the surface region

of the material. The growth of the hardened oxygen-containing case is accompanied by

building-up the residual stresses as a consequence of volume expansion. Therefore to assess

the stress state of the hardened-oxidized region, conventional non-destructive ex-situ X-ray

diffraction sin2ψ method was employed. The in-situ XRD sin2ψ was also utilized to

investigate the stress evolution during long-term air-oxidation consisting of heating,

isothermal holding, and cooling stages. Moreover, the relatively new minimally-

destructive incremental ring-core focused ion beam (FIB) milling was performed to

evaluate the oxidized surface strain release using digital image correlation (DIC) algorithm.

In the upcoming sections, an introduction to each technique as well as the experimental

approach in this study are provided.

Residual stress determination

52

Ex-situ (and in-situ) X-ray sin2ψ method

The principle of the lattice strain measurement (and stress analysis) by X-ray diffraction

is the determination of the stress state in the material by the measurement of the variation

of interplanar spacing (dhkl) in different directions, so-called sin2ψ method [24,25], which

was firstly reported by M. Joffe and F. Kirpitcheva in 1922 [26]. Stress is an extrinsic

property which cannot be directly measured by X-ray diffraction. Therefore, in order to

quantify the stresses, it is required to initially measure the atomic-scale strain changes and

the associated residual stress is determined from the elastic properties of the crystalline

material and the elastic interaction of differently oriented crystals. In other words, crystals

are used as internal strain gauges to monitor the effect of stresses on lattice spacing. The

response of a certain state of stress on the interplanar spacing ( hkld ) as compared to the

lattice spacing of a reference stress-free condition ( 0

hkld ) is different for different

measurement directions defined by azimuthal tilt (ψ) and in-plane rotation (φ) angles with

respect to the surface normal. This effect is reflected in a shift of the characteristic

diffraction peaks and the corresponding measurement gives strain [8,24].

In the determination of the RS for a surface-engineered material characterized by one-

dimensional into a flat sample using X-ray diffraction sin2ψ method, the stress distribution

is defined by the in-plane surface stresses (σ13= σ23= σ33=0) and the lattice strain ( ,

hkl

) is

calculated using the following equation [8]:

, 0 2 2 2

, 2 11 12 22 1 11 22

0

1cos sin 2 sin sin

2

hkl hkl

hkl hkl hkl

hkl

d dS S

d

3-4

where ,

hkl

is the strain experienced by the set of lattice planes {hkl} in a direction defined

by φ and ψ. ,

hkld and 0

hkld are the strained and strain-free interplanar spacing, respectively.

1

hklS and 2

1

2

hklS are the X-ray elastic constants (XECs) obtained from grain interaction

models10 [8,21].

10 In the case of this study Voigt assumption was used to calculate the XECs. This implies

that all crystals in the surface region of the oxidized BMGs experience the same strain.


53

The interplanar lattice spacing ( ,

hkld ) and lattice strain (, 0

,

0

hkl hkl

hkl

hkl

d d

d

) are computed

and plotted as a function of sin2ψ. For measuring strain-free direction (2

0sin ), the stress

state in Eq. 3-4 is simplified to a rotationally symmetric biaxial state of stress (σ11=σ22=σ//)

in the surface [8]:

2

, 2 / / 1 / /

1sin 2

2

hkl hkl hklS S 3-5

Plotting the lattice strain for various ψ angles and thereby the corresponding sin2ψ could

give a linear relation, where the stress can be obtained from the slope of the sin2ψ plot (see

for example a schematic representation provided in Fig. 3.7-b. From Eq. 3-5, 2

0sin is

given by 1

2

2

(1/ 2)

hkl

hkl

S

S . A value 0

hkld is calculated based on the interpolation in a

2

, sinhkld plot. The slope and intercept of the linear fit through the data points are

determined for measuring the stress values, i.e. σ11 and σ22 in the two surface directions

based on,

2

2 0

2

sin

hkl

hkl hkl

d

S d

. The standard error estimate of a linear fitting is also

considered. The definition of the lattice strain measurement for the overall state of stress in

the crystalline material is schematically provided in Fig. 3.7-a. The schematic illustration

(Fig 3.7-b) shows how a set of crystallographic planes oriented at a specified angle to the

surface are measured by tilting the sample so that the crystallographic planes are in a

position where they fulfil the Bragg condition. In other words, diffraction occurs from those

set of planes hkl which are parallel to the surface of the material (their diffraction vector is

perpendicular to the sample surface at an angle ψ to the surface).


54

Fig. 3.7. (a) Definition of lattice strain measurement directions, i.e. φ and ψ of the system of principal stresses

in the crystalline material surface with normal vector n. (b) Schematic representation of recording

( )hkld f for one specific hkl and a 2

, sinhkl

plot indicating a polycrystalline material under

homogeneous biaxial compressive stress.

In this study both ex-situ and in-situ X-ray diffraction sin2ψ measurements were utilized

for the determination of the macroscopic strain and residual stresses developed in the

surface of the oxidized BMGs:

Ex-situ lattice strain measurement

The ex-situ lattice strain analysis was conducted in parallel beam geometry using a Bruker

D8 X-ray diffractometer (Theta-2 Theta goniometer). The measurements were performed

at a specific 2θ range to include the main reflection of ZrO2. The step size (2θ) was chosen

0.03 and the counting time per step of 10 s. For this purpose, a series of diffractograms

were acquired at various ψ angles. The interplanar spacing is then calculated11 and its

variation is plotted as a function sin2ψ.

In-situ lattice strain measurement

X-ray diffraction lattice strain measurement was performed under the thermal cycle

comparable to the in-situ XRD oxidation for phase analysis (section 3.2.1) and the

investigation of the state of stress was monitored in-situ, applying the sin2ψ method. For

this analysis, the diffractometer was equipped with the X-ray source of Cu Kα radiation

11 The peak fitting of the diffractograms obtained from ex-situ (and in-situ) was performed

using Pseudo-Voigt function with the TOPAS software version (Bruker AXS)


55

(λ=0.154 nm)12 and the same setup used for the in-situ phase analysis was utilized. A ~

20×5×1 mm3 BMG sample was polished to 1 µm diamond and cleaned in an ultrasonic

bath with ethanol before applying the in-situ lattice strain determination. The sample was

heated up to 600K in the air in 20 K intervals; at each step, a series of diffractograms was

acquired at 5 ψ angles. The 2θ range 26°-47° has been chosen such that the various

reflections of oxide phases i.e. (011) plane of t-ZrO2, (111) plane of m-CuO and (111) plane

of c-Cu2O are present. The step size of 0.04 and the counting per step was 3 s and they were

selected such that the total acquisition time per stress measurement (5 diffractograms) was

about 2.2 h. After reaching 600 K, oxidation of the BMG is followed isothermally for 60

h. The state of stress is probed intermittently by applying a series of sin2ψ measurements;

the probing conditions are the same as applied during heating (i.e. one per ~ 2.2 h).

Eventually, the investigation is continued during cooling to 300 K, in steps of 50 K.

Incremental ring-core focused ion beam (FIB) milling and digital image

correlation (DIC)

The micron-scale incremental ring-core FIB-DIC milling method is a relatively new

technique to determine the in-plane residual stresses (and strains) [23]. This method was

firstly developed by Korsunsky et al. in 2009 [27] based on the macroscopic ring-core

method proposed by Keil in 1992 [28]. In this technique, local material removal within a

specific gauge volume at the surface region of the material using high energy (Ga+) ions

induces surface strain relief where the resulting strain changes can be tracked and evaluated

by DIC [29]. This method enables determining the residual stresses in amorphous materials

which cannot be measured using conventional diffraction techniques [30]. In the

experimental approach, a ring-core milling geometry, which provides a progressive

homogenous relaxation of strain in all directions at the surface region is mostly used

compared with other milling geometries [31,32]. The surface strain relaxation is then

digitally monitored by tracing deposited high-contrast features in high-resolution SEM

images obtained during the FIB milling process. The procedure of the FIB/DIC

methodology is as follows: initially, electron-beam assisted platinum (Pt) deposition of a

12 The X-ray tube source has been replaced to Cu-Kα for in-situ X-ray sin2ψ lattice strain

analysis to provide more information depth compared with Cr Kα.


56

speckle pattern13 on the region of interest at the sample surface is applied. The pattern

comprises a uniform but non-periodic array of Pt dots [33], which is used as a reference

pattern and enhances the contrast of SEM micrographs for DIC analysis [34]. An annular

trench is subsequently milled using the specific FIB milling parameters 14 around the

deposited pattern. It is noted that a sequence of high-resolution SEM images of the surface-

speckle pattern is acquired before and after each milling step. Following this, a series of

high contrast and high-resolution SEM micrographs acquired during FIB milling is then

digitally analyzed using DIC to measure strain changes at each milling cycle with respect

to the reference pattern in two directions at the surface [30,35]. In addition, to be sure that

reliable values are achieved consistently, several FIB milling processes should be

conducted at various sample surface sites. In the case of applying this method, some

important points are required to be taken into account to minimize the possible artifacts

introduced by FIB operation and process optimization [36]:

- The outer ring diameter should be at least > cylinder diameter (D) + (2×the

anticipated FIB milling depth (h) to avoid material re-deposition on to the pillar (or

cylinder). (cf. Fig. 3.8)

- A full stress relaxation (complete strain relief saturation) is obtained when the

milling depth is at least equal to the pillar (or cylinder) diameter. (h/D≥1)

- Adopting the outer to the inner path of the focused ion beam during incremental

milling to avoid material re-deposition on the cylinder.

- For analyzing the SEM images using DIC, drift, and noises in the SEM images need

to be minimal, otherwise, it will directly affect the strain results. For this purpose,

using auto-drift image correction function in a FIB/SEM microscope while

13 In order to obtain consistent and reliable SEM images for use in DIC analysis, the region

of the interest on a sample’s surface should be decorated by a dense, random and high

contrast surface-speckle pattern [33].

14 The specific operating parameters are specimen dependent. The current and dwell-time

of the focused ion-beam will directly affect the material removal rates and re-deposition

behaviour.


57

performing the process is highly recommended. Otherwise, manual drift correction

by the microscope operator using a fiducial may be required.

- The ion-beam drift must be monitored and, if required, corrected to prevent the

geometrical errors in the FIB milled area.

Fig. 3.8. Schematic illustration of the ring-core milling geometry adopted in this study where the cylinder

diameter and milling depth are annotated by D and h, respectively.

In this study, in addition to the XRD method, the novel incremental FIB/DIC milling which

is a semi-destructive technique, has been utilized as a complementary method for

measuring the strain relief in the surface region of the thermochemically surface-engineered

BMG at an extremely low pO2. For this purpose, a FEI Helios NanoLAB G3 CX dual-beam

at UOW Electron Microscopy Centre (EMC) has been used 15 , 16 . The microscope is

equipped with a Ga+ ion source and gas injection systems for C and Pt deposition. A

schematic illustration of the FIB-SEM setup is depicted in Fig. 3.10-a. For the stepwise

FIB milling process, the sample was tilted to 52˚, facing a FIB column direction

perpendicular to the sample surface. SEM images were acquired at 52º tilt to the electron-

beam, where the tilt correction has been activated to avoid a vertical compression of the

15 Automatization of the incremental ring-core FIB/DIC milling process was achieved by

applying the AutoScript 4 plugin on the instrument.

16 It is worth mentioning that several attempts were made for development and optimization

of FIB-DIC technique in the case of this study using an FEI Helios Nanolab 600 at the

center for electron microscopy (CEN) of DTU, but the manual drift correction of both

electron and ion beam was insufficient to compensate for the drift observed to obtain

reliable and reproducible results.


58

images. In the following sections, the experimental procedure required for conducting the

FIB-DIC method is provided17:

Surface-speckle pattern decoration

In order to apply DIC to SEM images, the sample surface features must provide sufficient

local contrast to enable accurate tracking of the surface displacement with DIC. If this local

surface contrast is not given by the sample itself, the region of interest on the sample surface

should be artificially decorated by a rigidly adhered pattern containing randomly distributed

speckles almost similar in size [37]. Several methods are available for applying the

stochastic patterns, among them the electron-beam (e-beam) lithography is of interest. The

e-beam lithography enables creating a consistently high-quality pattern and allows

optimizing the pattering parameters to enhance its resolution [38]. In this work, a uniform

and non-periodic speckle pattern [33] was applied on the sample surface using e-beam

assisted Pt deposition at 10 kV and 1.4 nA. For this purpose, the exact numerically

controlled e-beam positioning via stream files was used [39]. Fig. 3.9 provides an SEM

image of a 10×10 µm2 Pt-speckle pattern deposited on an oxidized BMG. The focus and

astigmatism require to be adjusted at high magnifications before applying the pattern.

Fig. 3.9. SEM image of a Pt-speckle pattern deposited on the sample surface using an e-beam at 10 kV

accelerating voltage and 1.4 nA electron beam current.

17 The process parameters used in this study were optimized for application of the method

to cubic-ZrO2 formed in the surface adjacent region of ZrCuAl-based BMG substrate.

While the procedure is generally valid for most types of materials, specific operating

parameters are sample dependent.


59

Optimization of SEM imaging

The essential preliminary prerequisite for a reliable surface strain measurement using DIC

is having a series of high magnification and high-resolution (high-pixel matrix) SEM

images. For this purpose, the image acquisition was set to the integration of 128 images at

a dwell time of 50 ns and 1536×1024 pixels resolution18. The high-resolution imaging was

performed at 10 kV accelerating voltage and 1.4 nA e-beam current using a through the

lens detector (TLD). To optimize 2D image correlation, high-contrast SEM images are

necessary. Therefore, before starting the FIB milling process, the brightness and contrast

of the SEM image were adjusted to evenly distribute the pixel intensities from 0 to ~90%

of the available histogram to keep a safety margin in case of contrast changes during

acquisition. The adjusted imaging parameters were kept unchanged until the end of the FIB

milling process.

FIB milling process and iterative high-resolution SEM imaging

During the FIB milling process, high energy Ga+ ions interact with the surface atoms of the

investigated material, and the energy is dissipated in several elastic and inelastic scattering

events which ultimately results in the physical removal of the target material [40]. The

sputtering is expected to be the main mechanism for material removal, where the impacted

surface and near-surface atoms are displaced from their original site. As a result, sputtering

could occur if these atoms have sufficient kinetic energy to overcome the surface binding

energy of the target material. Higher ion currents mill the material faster but have lower

resolution, and tend to increase the amount of re-deposited material [41]. In the current

work, the FIB intensity of 30 kV accelerating voltage and 230 pA ion beam current were

used for milling of cubic ZrO2. In order to have a homogenous strain relief in all directions

at the surface region, as stated previously a ring-core geometry was selected. The outer

diameter used in the surface strain analysis was set to at least 2 times of the milling depth

plus the cylinder diameter (2h+D). Each single-pass milling step was done at a dwell time

18 It should be specified that acquiring a single SEM image with high dwell time may result

in noise and drift of the images which can be problematic for use in post-processing by

DIC.


60

of 150 µs per pixel where the milling direction was set from the outer to the inner side of

the cylinder. After each milling step, a high-resolution SEM image was acquired iteratively

until the expected final milling depth was reached. Fig. 3.10-b presents a series of SEM

images obtained at the same magnification form the FIB milled area in the surface region

of the oxidized ZrCuAl-based BMG. The final FIB milling depth should be sufficient to

fully relieve the residual stresses in the selected region of the sample. In the case of this

study, a milling depth approximately equal to the pillar (or cylinder) diameter was

considered for a fully relaxed area. The milling depth could be affected by the re-deposition

of the material into the trench.

Fig. 3.10. (a) Schematic illustration of a ring-core FIB milling method. (b) A sequence of high-resolution

SEM micrographs acquired from the FIB milled area before and after each milling step at the same

magnification.

Post-processing by digital image correlation (DIC)

Digital image correlation (DIC) is an effective optical metrology method for computing in-

plane full-filed strain by comparing the images acquired from the surface of the sample

before and after deformation [42]. This method was first proposed and applied to the

measurement of macroscopic displacements by Sutton et al. in 1983 [43]. In principle, after

acquiring the images of the sample surface before and after each deformation state, the non-

contact DIC algorithm can digitally track the movement of each feature in the images and

compute the image pixel displacements in terms of strain [42,44]. In DIC, the region of

interest in the reference image is initially defined which is further subdivided into a grid of

facets. The displacement of each facet is measured to obtain the full-field surface strain.

The basic principle of DIC is the monitoring of the same pixel of two images, which gives

the integer transitional shift between them. To track the displacements of pixels, a square

reference subset is specified which facilitates finding match points. This approach allows


61

the accurate and rapid approximation of the relative displacement through the use of 2D

normalized cross-correlation [23,29]. It compares the intensity of pixels in the reference

facets with the pixels in the strained facets and cross-correlates the intensities to determine

an average shift of the facet position corresponds to its center point (DIC markers). The

global displacement fields between the reference and subsequent images are then quantified

by evaluating the shifts determined at each DIC marker location [29]. Here, it is noted that

it is difficult for the DIC algorithm to track the matched point using a single pixel. The

size of the subsets requires to be possibly small to maximize the spatial resolution, but large

enough to capture sufficient contrast features for reliable displacement measurement. It is

worth mentioning that the key point for DIC post-processing is using high-resolution well-

focused and low-noise images with high contrast which are required to be consistent over

several sequential images. In this study, after collecting the acquired SEM images from

the FIB milled area, the DIC post-processing was carried out in the freely available DIC

MATLAB script19. A detailed utilization of DIC analysis for tracking the surface strain

relaxation during the FIB milling process is comprehensively provided and discussed in the

result chapter 8 of this dissertation.

19 The DIC algorithm is based on the function cpcorr.m of the MATLAB Image Processing

Toolbox and a first implementation was done by C. Eberl, D.S. Gianola and S. Bundschuch

in 2010. (https://www.mathworks.com/matlabcentral/fileexchange/12413-digital-image-

correlation-and-tracking). The updated version of DIC MATLAB code developed by F.

Niessen in 2019 has been used in the current study.

https://www.mathworks.com/matlabcentral/fileexchange/12413-digital-image-correlation-and-tracking

https://www.mathworks.com/matlabcentral/fileexchange/12413-digital-image-correlation-and-tracking


62

References

[1] R. f. Speyer, Thermal analysis of materials, Taylor & Francis, 1994.

[2] W.C. Röntgen, Ueber eine neue Art von Strahlen, Ann. Phys. 300 (1895) 12–17.

[3] R. Jenkins, R.L. Snyder, Introduction to X‐ray Powder Diffractometry, John Wiley

& Sons, 1996.

[4] A.A. Bunaciu, E. gabriela Udriştioiu, H.Y. Aboul-Enein, X-Ray Diffraction:

Instrumentation and Applications, Crit. Rev. Anal. Chem. 45 (2015) 289–299.

[5] W.H. Bragg, W.L. Bragg, The Reflection of X-rays by Crystals, 17 (1913) 428–438.

[6] F. Niessen, Supermartensitic Stainless Steels PhD Thesis, 2018.

[7] E.J. Mittemeijer, P. Scardi, Diffrcation analysis of the microstructure of materials,

Springer Berlin Heidelberg, 2004.

[8] V. Hauk, Structural and residual stress analysis by non-destructive methods:

Evaluation-Application-Assessment, 1997.

[9] G.F. Vander Voort, ed., Metallography and Microstructures Handbook, ASM

International, 2004.

[10] Y. Leng, Materials Characterization: Introduction to Microscopic and Spectroscopic

Methods, Wiley, 2013.

[11] N.E. Dowling, Mechanical behavior of Materials, 4th ed., Pearson Education, 2012.

[12] ASTM E384 - 17, Standard Test Method for Microindentation Hardness of

Materials, ASTM Int. (2017) 1–40.

[13] L. Reimer, Scanning electron microscopy: physics of image formation and

microanalysis, Second edi, Springer Berlin Heidelberg, 1998.

[14] W. Zhou, R. Apkarian, Z.L. Wang, D. Joy, Fundamentals of scanning electron

microscopy (SEM), in: Scanning Microsc. Nanotechnol. Tech. Appl., Springer

Science+Business Media New York, 2006: pp. 1–40.

[15] R.F. Egerton, The Scanning electron microscope, Phys. Princ. Electron Microsc.

(2010) 125–153.

[16] J.I. Goldstein, D.E. Newbury, J.R. Michael, N.W.M. Ritchie, J.H.J. Scott, D.C. Joy,

Scanning electron microscopy and x-ray microanalysis, Third, Springer

Science+Business Media New York, 2003.

[17] Hawkes, Peter W., J.C.H. Spence, Springer Handbook of Microscopy, 2019.

[18] B.J. Inkson, Scanning Electron Microscopy (SEM) and Transmission Electron

Microscopy (TEM) for Materials Characterization, Elsevier Ltd, 2016.


63

[19] S.K. Sharma, D.S. Verma, L.U. Khan, S. Kumar, S.B. Khan, Handbook of Materials

Characterization, 2018.

[20] P.J. Goodhew, J. Humphreys, R. Beanland, Electron microscopy and analysis,

Taylor & Francis, 2001.

[21] I.C. Noyan, J.B. Cohen, The nature of residual stress and its measurement, in:

Springer Science+Business Media New York, 1982.

[22] P.J. Withers, H.K.D.H. Bhadeshia, Residual stress part 1 - Measurement techniques,

Mater. Sci. Technol. 17 (2001) 355–365.

[23] Alexander M. Korsunsky, A Teaching Essay on Residual Stresses and Eigenstrains,

2017.

[24] American Society for Matals Handbook, Materials Characterization, Volume 10,

ASM International, 1986.

[25] M.R. James, J.B. Cohen, Measurement of Residual Stresses By X-Ray Diffraction

Techniques., ACADEMIC PRESS, INC., 1980.

[26] A.F. Joffe, M.V. Kirpitcheva, Röntgenograms of strained crystals, London,

Edinburgh, Dublin Philos. Mag. J. Sci. 43 (1922) 204–206.

[27] A.M. Korsunsky, M. Sebastiani, E. Bemporad, Focused ion beam ring drilling for

residual stress evaluation, Mater. Lett. 63 (2009) 1961–1963.

[28] S. Keil, Experimental Determination of Residual Stresses With the Ring‐Core

Method and an on‐Line Measuring System, Exp. Tech. 16 (1992) 17–24.

[29] A.J.G. Lunt, A.M. Korsunsky, A review of micro-scale focused ion beam milling

and digital image correlation analysis for residual stress evaluation and error

estimation, Surf. Coatings Technol. 283 (2015) 373–388.

[30] A.J.G. Lunt, N. Baimpas, E. Salvati, I.P. Dolbnya, T. Sui, S. Ying, H. Zhang, A.K.

Kleppe, J. Dluhoš, A.M. Korsunsky, A state-of-the-art review of micron-scale

spatially resolved residual stress analysis by FIB-DIC ring-core milling and other

techniques, J. Strain Anal. Eng. Des. 50 (2015) 426–444.

[31] A.M. Korsunsky, M. Sebastiani, E. Bemporad, Residual stress evaluation at the

micrometer scale: Analysis of thin coatings by FIB milling and digital image

correlation, Surf. Coatings Technol. 205 (2010) 2393–2403.

[32] E. Bemporad, M. Brisotto, L.E. Depero, M. Gelfi, A.M. Korsunsky, A.J.G. Lunt, M.

Sebastiani, A critical comparison between XRD and FIB residual stress

measurement techniques in thin films, Thin Solid Films. 572 (2014) 224–231.

[33] F. Di Gioacchino, W.J. Clegg, Mapping deformation in small-scale testing, Acta

Mater. 78 (2014) 103–113.

[34] H. Zhang, E. Salvati, S. Ying, A. Lunt, T. Sui, A.M. Korsunsky, Effect of Pt


64

deposition on digital image correlation analysis for residual stress measurement

using FIB-DIC ringcore method, Lect. Notes Eng. Comput. Sci. 2224 (2016) 838–

842.

[35] A.M. Korsunsky, E. Salvati, A.G.J. Lunt, T. Sui, M.Z. Mughal, R. Daniel, J. Keckes,

E. Bemporad, M. Sebastiani, Nanoscale residual stress depth profiling by Focused

Ion Beam milling and eigenstrain analysis, Mater. Des. 145 (2018) 55–64.

[36] X. Song, K.B. Yeap, J. Zhu, J. Belnoue, M. Sebastiani, E. Bemporad, K. Zeng, A.M.

Korsunsky, Residual stress measurement in thin films at sub-micron scale using

Focused Ion Beam milling and imaging, Thin Solid Films. 520 (2012) 2073–2076.

[37] Y.L. Dong, B. Pan, A Review of Speckle Pattern Fabrication and Assessment for

Digital Image Correlation, Exp. Mech. 57 (2017) 1161–1181.

[38] A.D. Kammers, S. Daly, Small-scale patterning methods for digital image

correlation under scanning electron microscopy, Meas. Sci. Technol. 22 (2011).

[39] F. Niessen, M.J.B. Nancarrow, Computer-aided manufacturing and focused ion

beam technology enable machining of complex micro- And nano-structures,

Nanotechnology. 30 (2019).

[40] L.A. Giannuzzi, B.I. Prenitzer, B.W. Kempshall, Ion-solid interactions, in: Introd. to

Focus. Ion Beams Instrumentation, Theory, Tech. Pract., Springer US, 2005: pp. 13–

52.

[41] L.A. Giannuzzi, F.A. Stevie, Introduction to focused ion beams: Instrumentation,

theory, techniques and practice, Springer US, 2005.

[42] B. Pan, K. Qian, H. Xie, A. Asundi, Two-dimensional digital image correlation for

in-plane displacement and strain measurement: A review, Meas. Sci. Technol. 20

(2009).

[43] M. Sutton, W. Wolters, W. Peters, W. Ranson, S. McNeill, Determination of

displacements using an improved digital correlation method, Image Vis. Comput. 1

(1983) 133–139.

[44] P.C. Hung, A.S. Voloshin, In-plane strain measurement by digital image correlation,

J. Brazilian Soc. Mech. Sci. Eng. 25 (2003) 215–221.

Summary of results

65

4 Summary of results

A summary of the main results of the current Ph.D. work is given in this chapter. For a

detailed and comprehensive discussion on key research findings, the readers are referred to

the experimental result chapters 5 to 8. The graphical abstract of each study is also

presented.

4.1 Manuscript I

Self-repair by stress-induced diffusion of the noble element during oxidation of

Zr48Cu36Al8Ag8 bulk metallic glass

The present study addresses an investigation to determine the oxidation mechanism of

Zr48Cu36Al8Ag8 BMG during long-term treatment at 600 K (~100 K below the glass

transition temperature, Tg) in air. The surface oxidized region was characterized using in-

situ X-ray diffraction and advanced electron microscopy techniques. The oxide region is

found to consist of two different zones, i.e. an outer oxidation zone (OOZ), comprised of a

stratified distribution of copper oxide, metallic silver, and nano-crystalline zirconia, and an

internal oxidation zone (IOZ) consisting of mainly zirconia and self-healed micro-cracks

(or decorated shear bands) enriched with copper inclined to the surface. The proposed

oxidation mechanism is as follows: the nano-crystalline ZrO2 formation in the IOZ leads to

volume expansion and thus compressive stress which drives metallic copper and silver

towards the surface where they segregate and crystallize, Cu is readily oxidized; Ag does

not oxidize in air at 600K. If internal surfaces develop by cracking, then these provide a

location for segregation and crystallization of Cu (and Ag), which repairs the crack (self-

healing). In the OOZ, cracks develop parallel to the surface as a consequence of

compressive stress (onset of spallation). Cu and Ag also segregate/crystallizes at these

Manuscript II

66

surfaces and Cu is thereafter oxidized. This gives the stratified microstructure, which grows

progressively into the IOZ on prolonged oxidation.

4.2 Manuscript II

Surface hardening by gaseous oxidizing of (Zr55Cu30Al10Ni5)98Er2 bulk-metallic glass

This study investigates the effect of gaseous oxidizing treatment (<Tg) on surface hardness

of (Zr55Cu30Al10Ni5)98Er2 BMG in a controlled oxidizing atmosphere providing extremely

low and high oxygen partial pressure (pO2). Several techniques including X-ray diffraction,

(electron) microscopy combined with energy dispersive spectroscopy and micro-hardness

test are utilized to characterize the oxidized surface zones. The results demonstrate that

applying oxidizing treatment at extremely high pO2 (≈1648 bar) leads to the development

of two different zones, i.e. an OOZ comprised of CuOx network due to the outward

diffusion of Cu and an IOZ consisting of mainly nano-crystalline tetragonal ZrO2. In the

gas, with an extremely low pO2 (≈10-26 bar), only the IOZ grows where the crystallization

of the BMG is also detected. The micro-hardness results of thermochemically oxidized

BMG show that a hardened case in IOZ can be obtained where its value reaches

approximately ~1200 HV adjacent to the surface.

Summary of results

67

4.3 Manuscript III

Stress, stress relaxation and self-healing in (Zr55Cu30Al10Ni5)98Er2 bulk metallic glass

during air-oxidation

In this study, the reaction of oxygen at a (Zr55Cu30Al10Ni5)98Er2 BMG surface was followed

in-situ using X-ray diffraction analysis to elucidate the correlation between the surface

microstructure evolution and the concomitant residual stresses. When the BMG is exposed

to the atmospheric conditions at temperatures below Tg, oxidation takes place which

induces a transformation of the amorphous state into crystalline oxide compounds in the

surface region. Ex-situ microscopical investigations show that the oxygen-containing

surface zones consist of an IOZ and an OOZ comprised of mostly nano-crystalline

tetragonal ZrO2 and Cu-based oxides, respectively. Moreover, in the IOZ, the development

of micro-cracks enriched with Cu/CuOx is observed. In-situ investigation of the evolution

of the state of stress in three oxide phases (tetragonal ZrO2, monoclinic CuO, and cubic

Cu2O) developed in the surface region of the investigated BMG during long-term sub-Tg

exposure to the air-atmospheric condition is applied using the XRD sin2ψ method. The

results show that the dissolution of oxygen leads to an expansion of the BMG surface by

the formation of ZrO2, and thereby the development of compressive residual stresses of ~

-1.4 GPa in this region. This is accompanied by Cu-redistribution and transport of Cu to

the surface, which tends to alleviate the growth stresses in the IOZ. Moreover, stress relief

occurs as a result of perpendicular crack initiation due to the compensating tensile stresses

in the BMG. Therefore, stresses induced by t-ZrO2 formation are largely relaxed from ~ -

1400 MPa to ~ -400 MPa during oxidation. The formation of crack surfaces provides new

Manuscript IV

68

locations for the crystallization and segregation of Cu and consequently self-healing of the

crack. Subsequent conversion of Cu to CuOx at the surface develops the OOZ.

4.4 Manuscript IV

Strain, stress, and stress relaxation in oxidized ZrCuAl-based bulk metallic glass

This study investigates the stress (and strain) introduced as a consequence of ZrO2 (Al2O3

and TiO2) formation on thermochemically oxidized Zr51.3Cu31.3Al8.5Ni4Ti4.9 BMG. For this

purpose, two independent techniques, i.e., X-ray diffraction sin2ψ and incremental ring-

core focused ion beam (FIB) milling methods are utilized. The BMG was oxidized in a

controlled gaseous atmosphere imposing an extremely low pO2 (10-41 bar) at 600 K for 60

h. The sin2ψ analysis is conducted on the (111) reflection of the crystalline cubic-ZrO2 peak

where it reveals the existence of compressive stress in ZrO2 of -1.5 GPa, corresponding to

macro-strain of ~ 0.45%. Surface strain relief monitored in high-resolution SEM imaging

during gradual ring-milling is determined by digital image correlation (DIC) and also

indicated the occurrence of compressive residual stresses of -1.4 GPa, corresponding to

macro-strain of ~ 0.5% in the internal oxidation zone. Furthermore, the stress relaxation

Summary of results

69

mechanism is explained in relation to the observed microstructural features, i.e. (i)

development of shear bands as a consequence of plastic deformation of the BMG and (ii)

redistribution and outward diffusion of Cu (and Ni) to the surface and decoration of shear

bands/corner cracks with Cu (self-healing phenomenon).

Manuscript IV

70

Manuscript I

71

5 Manuscript I

Self-repair by stress-induced diffusion of noble elements during of

oxidation of Zr48Cu36Al8Ag8 bulk metallic glass1

Saber Haratian, Flemming B. Grumsen, Matteo Villa, Thomas L. Christiansen, Marcel A.J.

Somers

Materials and Surface Engineering Section, Mechanical Engineering Department,

Technical University of Denmark, Produktionstorvet, Building 425, 2800 Kgs. Lyngby,

Denmark

Abstract

Long-term oxidation behavior of biocompatible Zr48Cu36Al8Ag8 bulk metallic glass

(BMG) under atmospheric conditions at 600 K (well below Tg) was investigated. The

microstructural evolution and surface morphology of the oxidized zone developing during

oxidation of the Ag-containing ZrCu-based BMG was determined, applying in-situ X-ray

diffraction and advanced electron microscopy techniques. The results demonstrate that two

oxide zones develop during atmospheric oxidation: an outer oxide zone (OOZ), consisting

of a stratified distribution of copper oxide (CuOx) regions, metallic silver, and zirconia, and

an inner oxide zone (IOZ) mainly consisting of tetragonal ZrO2 showing copper

enrichments along lines inclined to the surface. The formation of CuOx and Ag in the OOZ

is ascribed to the outward diffusion of Cu and Ag driven by the compressive stress that

develops upon internal oxidation of Zr (and Al) to t-ZrO2 (and Al2O3). The presence of the

stratified microstructure in OOZ and the Cu enrichments in the IOZ are ascribed to micro-

cracks resulting from the compressive stresses induced by volume expansion associated

with ZrO2 formation. Segregation of metallic elements towards these oxidation-generated

free micro-crack surfaces “repairs” the cracks and can consistently explain the observed

microstructural features.

1 Published work: S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, M. A.J. Somers.

“Self-repair by stress-induced diffusion of noble elements during oxidation of

Zr48Cu36Al8Ag8 bulk metallic glass,” Scr. Mater., vol. 164, pp. 126–129, 2019. The format

of the published article was adapted to the format of the Ph.D. thesis.

Introduction

72

5.1 Introduction

Bulk amorphous alloys (BAAs), consisting of at least two principal components and

additional alloying elements, have been extensively investigated over the last few decades

[1–3]. Large effort has been devoted to the development of metallic glassy systems,

motivated by their fundamental outstanding properties as high specific strength, high elastic

limit, good corrosion and wear resistance, low Young’s modulus, excellent surface finish

and super-plasticity in the super-cooled liquid region [4–6]. Among metallic glasses, in

particular the Zr-based bulk glassy alloys possess a desirable combination of properties and

high glass-forming ability (GFA), which makes them attractive for potential exploitation

in various structural engineering applications [4,7,8], including biomedical parts and tools,

sporting equipment, aerospace precision counterparts, automotive parts and jewelry.

Hitherto, most studies have focused on enhancing the plasticity [9–12], thermal stability

[13–14] and GFA [15–16] of metallic glasses to improve their functional capabilities. Zr-

based bulk glassy systems free of toxic Be and allergenic Ni are expected to be utilized for

biomedical applications [17–25].

Surface engineering of Zr-based bulk glassy systems is hypothesized to be possible by

incorporating oxygen at (moderately) high temperatures, due to the large solid solubility of

O in Zr. So far the oxidation behavior of Zr-based BMGs has been investigated, albeit not

for the purpose of surface engineering, but rather to investigate the oxidation behavior.

Several published investigations [26–31] reported the oxidation behavior of Zr-based

BMGs over the temperature range of their super-cooled liquid region (ΔTx). Zhang et al.

[32] studied the oxidation behavior of the Zr55Cu30Al10Ni5 BMG within the temperature

range 693-743 K for different durations in air. They found that the scales formed during

oxidation consist mostly of tetragonal and monoclinic ZrO2 along with a small fraction of

Al2O3. Cu and CuO particles were observed in the outer oxide scale. Lim et al. [33]

investigated how the oxidation takes place during continuous heating from room

temperature to well above the crystallization temperature (Tx) of ZrCu-based BMGs.

Specifically, they investigated the effect of Al as an alloying element and its oxidation

resistance in the supercooled liquid region. It has been stated that in the case of the binary

system of Zr50Cu50, the oxide layer consists of monoclinic ZrO2 and metallic copper.

Contrary to this, the oxide layers that develop on a ternary Zr46Cu46Al8 alloy contain

tetragonal and monoclinic ZrO2, Al2O3 and a small amount of metallic Cu, suggesting that

Manuscript I

73

the growth kinetics of the oxide scale is governed by inward diffusion of oxygen for

oxidation of Zr and Al and outward transport of copper, which subsequently crystallizes at

the surface. Here, it is noted that annealing metallic glasses below ΔTx could have a

detrimental effect on the fracture toughness due to the annihilation of free volume in the

glassy alloy [34].

The present investigation is an experimental study of long-term (60 h) oxidation in air of a

Zr-based BMG containing noble metals, Ag and Cu, below its glass transition temperature

using in-situ X-ray diffraction analysis and electron microscopy. In particular, the

precipitation behavior of metallic silver in the topmost oxide zones and the formation of

copper oxide were investigated.

5.2 Materials and methods

The material under investigation was ZrCu-based bulk amorphous alloy with nominal

composition Zr48Cu36Al8Ag8. The material was synthesized using vacuum arc melting of

high purity (99.9 wt. %) elemental constituents under a Ti-gettered high-purity argon

atmosphere. The produced ingots of the alloy were re-melted four times to ensure

compositional homogeneity. Homogenous 2×10×60 mm3 plate ingots were cast in a copper

mold (Rapid Quench Machine System VF-RQT50, Makabe Co. Ltd. Japan).

The thermal stability and crystallization behavior of the studied BMG was determined with

differential thermal analysis (DTA) in a Netzsch STA 449C simultaneous thermal analyzer

under a flow of argon (flow rate of 50 cm3/min; purity 99.999%). A sample of about 40 mg

was positioned in an alumina (Al2O3) crucible and isochronally heated from 298 K to 1273

K at 10 K/min. The glass transition temperature (Tg), and the crystallization temperatures

(Tx) were measured to 700 K and 770, respectively.

The oxidation was followed in-situ in a Bruker D8 Discover X-ray diffractometer. The in-

situ XRD oxidation experiment was applied to a sample of dimensions 3×1×0.5 mm3 which

was heated up to 600 K in 20 K intervals; for each step a diffractogram was recorded. The

heating rate in-between steps was 60 K/min. Upon reaching 600 K, the sample was kept at

this temperature for 60 hr. The cooling to ambient temperature was done in temperature

steps of 50 K. The cooling rate in-between the steps was 60 K/min. X-ray diffractometry

Results and interpretation

74

was performed with Cr Kα radiation (λ=0.229 nm) in a parallel beam geometry for the 2θ-

range 42°-75°; the step size was 0.06° 2 and the counting time per step was 3 s. The

temperature accuracy during in-situ XRD experiment was ± 10 K. Before oxidation

experiment, BMG sample was carefully polished to 1 µm diamond and then cleaned with

ethanol.

Cross-sections of the oxidized BMG were investigated in a Neophot 32 (Zeiss, Jena) light

optical microscope (LOM) and a Merlin scanning electron microscope (SEM, Carl Zeiss)

equipped with a Bruker Xflash 6ǀ60 energy dispersive (EDS) detector. The surface

morphology of the oxidized sample was characterized in a FEI Helios Nanolab 600 field

emission scanning electron microscope (FEG-SEM), equipped with a focused ion beam

(FIB). The sample oxidized in the in-situ XRD experiment was investigated in a JEOL

3000F transmission electron microscope (TEM) coupled with EDS. The preparation of the

TEM lamella was performed in the FEI Helios Nanolab 600, using the available Ga+ ion

source of the focused ion beam (FIB). Pt-deposition was applied to protect the sample

surface and an Omniprobe micro-manipulator was used for transfer of the electron

transparent sample to a Cu grid.

5.3 Results and interpretation

In-situ X-ray diffraction

The in-situ XRD results are shown in Fig. 5.1. Initially, in the heating stage, the sample is

amorphous as evidenced by the diffuse peak which narrows with increasing temperature

(dotted line). The presence of faint diffraction peaks marked by arrows and present from

room temperature are the supporting Al2O3 sample holder rods, which is unavoidable in the

test set-up for the small sample. On heating, tetragonal ZrO2 formed at a temperature as

low as ~450 K, which is attributed to the high affinity of Zr for O. On continued heating

and oxidation the intensity of ZrO2 is reduced while Cu2O, CuO and Ag become visible.

This suggests that the latter phases appear at the surface on top of the ZrO2. Also, metallic

Cu is visible, but this peak soon vanishes while Cu2O and CuO peaks appear. On continued

oxidation the intensity of CuO appears to increase while that of Cu2O decreases with

oxidation time, indicating that CuO is the most stable of the Cu oxides.

Manuscript I

75

In the cooling stage (see top of Fig. 5.1), the peak positions of most phases shift to higher

Bragg angles as a consequence of thermal contraction. The peak positions of the Cu2O

phase, however shift to lower Bragg positions, indicating the thermal contraction is

counteracted by (additional) compressive stresses as compared to the other phases.

Fig. 5.1. XRD results of the Zr48Cu36Al8Ag8 BMG during isothermal heating at 600 K for 60 hr and

isochronal cooling in air. The 2D plot gives the diffractograms vs time, including heating, isothermal

and cooling stages. The square root intensity is indicated by the color according the scale on the right.

Diffraction lines indicated by arrows result from the Al2O3 rods to mount the sample and could be seen

as a reference.

Electron microscopical characterization of the oxide scale

A secondary electron (SE) micrograph of the surface morphology of the oxidized BMG

treated at 600 K for 60 hr is given in Fig. 5.2-a. The micrograph indicates grey irregularly

distributed granular and hemispherical porous particles of different sizes, covering the

entire surface. Needle-like oxide features (white) grown on top of oxide precipitates are

observed, too. Moreover, small mushroom-like particles (indicated by arrows) are unevenly

distributed between the large precipitates. In order to determine the chemical composition

of the particles formed on the surface, energy dispersive spectroscopy (EDS) point analyses

were performed at selected locations. The results show that the surface region is enriched


76

in Cu, Ag and O. Additionally, according to the estimated2 elemental concentration values,

the mushroom-like particles are enriched in metallic silver. These results are consistent

with the XRD results where the development of non-oxidized Ag and Cu-oxides was

observed. Apparently, metallic Ag is heterogeneously distributed over the surface and also

present underneath the copper oxide regions. A cross-sectional light optical image of the

investigated oxidized BMG is provided in Fig. 5.2-b. The oxide scale can be subdivided in

an inner and an outer oxide zone (IOZ and OOZ, respectively); the transition from the IOZ

to the un-oxidized BMG is practically parallel to the surface. The various contrasts in the

two zones indicate the existence of several phases. Arrows in the IOZ in Figs. 5.2-b and

5.2-d mark the occurrence of Cu-colored (red in Fig. 5.2-b) lines, inclined with respect to

the surface. These lines are interpreted as cracks which were filled with pure Cu, suggesting

that the cracks developed at the oxidation temperature and that Cu diffused to these newly

created free surfaces, filling out the cracks, thus providing a self-healing mechanism.

Alternatively, the inclined features could be interpreted as decorated shear bands. The

presence of Cu is confirmed by EDS analysis (Fig. 5.2-d). The bright white (highest BSE

yield) dots in the OOZ as observed in Fig. 5.2-b and 5.2-c are interpreted as discontinuous

Ag layers as is confirmed by EDS in Fig. 5.2-d. The IOZ is depleted in Cu and Ag and is

enriched in O (Fig. 5.2-d). The sharp IOZ/BMG transition indicates that the IOZ develops

under the influence of inward diffusion of O and is accompanied by outward diffusion of

Cu and Ag, which accumulate in the OOZ (Fig. 5.2-d). The OOZ displays a stratified

microstructure and contains alternating layers of Cu-oxides, Ag and ZrO2. This could be

interpreted as a repeating Cu/Ag redistribution during oxidation of Zr to ZrO2. The OOZ

displays a stratified microstructure and contains alternating layers of Cu-oxides, Ag and

ZrO2. This could be interpreted as a repeating Cu/Ag redistribution during oxidation of Zr

to ZrO2. The presence of porosity in OOZ aligned parallel to the surface (Fig. 5.2-c)

indicates that cracks parallel to the surface have been present. It is suggested that the

2 EDS analyses can only be approximate if light elements are present and for a

heterogeneous microstructure, both along and perpendicular to the surface.

Manuscript I

77

volume expansion associated with ZrO2 formation and the associated compressive stress3

causes cracks parallel to the surface and induces a redistribution of Cu and Ag by stress-

induced diffusion.

Fig. 5.2. (a) SE micrograph of the top surface morphology of oxidized Zr48Cu36Al8Ag8 BMG; (b) cross-

sectional light optical microscopy; (c) Cross-sectional BSE micrograph of the oxide zones developed on

Zr48Cu36Al8Ag8 BMG; (d) EDS mapping of selected area (dotted pink rectangular in b.) of the oxidized

metallic glassy system.

TEM investigation was carried out to obtain more detailed information on the phase

distribution in the OOZ. For this purpose the in-situ XRD oxidized sample was used. The

TEM study was performed on a 4 µm long lamella extracted from the oxidized sample

surface. Investigation of the OOZ in the direction perpendicular to the surface in the TEM

revealed a very fine structural distribution of the oxide phases. An overview of the lamella

showing different regions with different contrast is provided in Fig. 5.3-a. TEM bright-field

3 The development of compressive stresses as a consequence of internal ZrO2 formation on

oxidizing an ZrCuAl- based BMG was recently demonstrated by X-ray diffraction and

incremental ring-core FIB milling method and will be presented in a forthcoming

manuscript [35].


78

imaging combined with EDS as given in Figs. 5.3-b and d for selected areas in Fig. 5.3-a

shows that the outer oxide region is a mixture of different phases consisting of copper

oxide, metallic silver, and tetragonal zirconia, in agreement with the results in Figs. 1-2.

The chemical compositions obtained from the various regions in Fig. 5.3 are consistent

with copper oxide, silver and mixed oxide zones (I, II and III). Underneath the copper

oxide-rich zone, a discontinuous sandwich-like distribution of metallic Ag-rich grains is

observed. At various locations within the Ag grains (annealing) twins are identified. The

mixed oxide zones consist predominantly of tetragonal ZrO2, and are located adjacent to

grains of silver and copper oxide. At various locations pores are found, consistent with the

porosities in Fig. 5.2-c. Besides the three main regions identified in the outer oxide zones,

some grains, designated as IV and V, appear to contain Ag, Cu, and O, indicating that

metallic silver grains are surrounded by copper oxide grains.

Fig. 5.3. (a) TEM bright field micrograph of the OOZ obtained from the oxidized Zr48Cu36Al8Ag8 BMG

during in-situ XRD experiment at 600 K for 60 hr in the air from the depth of ~4 µm. Magnified views of the

selected area 1-3 in (a) are given in (b) - (d), respectively, together EDS point analyses indicating the

approximate composition of the regions.

Manuscript I

79

Selected area electron diffraction (SAED) was applied to identify the crystallography of

the phases obtained from position 1, 2 and 3; the results are given in Fig. 5.4. The SAED

patterns confirm the existence of CuO (monoclinic), metallic Ag (fcc) and ZrO2 (tetragonal)

which is in excellent agreement with the results acquired from XRD experiments (and Fig.

5.1). Whereas for CuO and Ag SAED patterns for individual grains can be obtained, ring-

type patterns where acquired for ZrO2, indicating nano-crystallinity of the ZrO2 grains.

Interestingly, EDS revealed the presence of Al in the ZrO2 areas, but no diffraction spots

of Al oxides were visible in the SAED. This result is consistent with the XRD data where

no Al oxides were observed. It is anticipated that in IOZ, Al is bounded to O in an

amorphous Al2O3 structure or dissolved in ZrO2.

Fig. 5.4. (a) Marked positions (1, 2 and 3) in the outer oxide film on the Zr48Cu36Al8Ag8 BMG. Typical

selected area electron diffraction (SAED) pattern of position (b) 1; (c) 2, (d) 3.

5.4 Discussion

The results of the microstructural analysis in Sections 5.3.1 and 5.3.2 demonstrate the

subdivision in an outer and an inner oxidation zone. The IOZ consists of ZrO2, is depleted

in Ag and Cu and contains Cu enrichments along lines inclined with respect to the surface,

which are interpreted as self-healed cracks. The IOZ develops under the influence of inward

Discussion

80

diffusion of oxygen. The OOZ has a microstructure that is stratified along the surface,

containing Cu oxide, nano-crystalline ZrO2 and metallic Ag.

In an early stage of oxidation metallic Cu is observed (Fig. 5.1). This could be Cu along

small cracks but could also be Cu at the surface before it is oxidized to Cu2O and later CuO.

From the results obtained, the following oxidation mechanism emerges. The strong affinity

of Zr for O leads to the development of nano-crystalline ZrO2 grains. Since Cu and Ag do

not dissolve in ZrO2, these elements need to be redistributed. The compressive stresses that

develop as a consequence of the volume expansion associated with the volume increase by

the conversion of Zr into ZrO2 [35] provide the driving force for outward diffusion of Cu

and Ag, similary as described in [36]. The nano-crystallinity of ZrO2 is likely to enable

diffusion of the metallic elements along the abundant matrix in-between ZrO2 nanoparticles

or, if ZrO2 is the only remaining phase, along grain boundaries. Since Cu and Ag have

limited mutual solubility, Cu and Ag crystallize separately at the surface and Cu is readily

oxidized in air at 600 K. The presence of un-oxidized metallic Cu regions in the IOZ along

lines inclined with the surface, indicate that the activity of oxygen (i.e. the apparent

equilibrium partial pressure) is too low in this region to oxidize Cu. The inclination of the

Cu-enriched lines indicates that the cracks were formed by a shearing mechanism.

Apparently, the compressive stresses in the IOZ due to volume expansion become so high

that they can no longer be accommodated elastically. Once a crack is formed, the free

(internal) surface thus created provides a location for Cu (and Ag) segregation.

Crystallization of the metallic element(s) will fill out and “repair” the crack. It cannot be

excluded that these regions in the IOZ also contain small amounts of Ag; this was not

investigated in detail and EDS (Fig. 5.2-d) was inconclusive in this respect.

The stratified OOZ is anticipated to have developed as follows. Compressive stress

imposed by the IOZ could lead to local cracks parallel to the surface, thus initiating

spallation. Thereby a new (internal) surface is created where Cu and Ag, arriving from the

interior by stress-induced outward diffusion, can segregate and crystallize and largely

“repair” the crack. The BSE micrograph in Fig. 5.2-c (black spots) and TEM BF

micrographs Fig. 5.3 a-c (white areas) shows porosity along lines parallel to the surface,

which could corroborate this interpretation. Subsequently, in the OOZ Cu oxidizes and

forms Cu2O and CuO, because the oxygen activity is higher than in the IOZ. Repetition of

Manuscript I

81

this mechanism of a local initiation of spallation induced by compressive stress imposed

by the inwardly growing IOZ, Cu/Ag segregation and Cu oxidation leads to the stratified

microstructure, which progressively grows into the IOZ on continued oxidation. The

proposed oxidation mechanism is schematically represented in Fig. 5.5.

Fig. 5.5. Schematic representation of the oxidation mechanism for Zr48Cu36Al8Ag8 BMG during isothermal

heating at 600 K for 60 hr under atmospheric conditions.

5.5 Conclusion

In summary, the long-term oxidation behavior of Zr48Cu36Al8Ag8 BMG at 600 K for 60 hr

under atmospheric conditions was investigated using in-situ XRD and advanced electron

microscopy techniques. The developing oxide scale can be subdivided in an outer (OOZ)

and an inner oxide zone (IOZ). In the initial stage of oxidation, the dissolution of oxygen

induces crystallization of tetragonal ZrO2 at a temperature as low as ~450 K, which is about

250 K below the glass transition temperature. Along with the development of nano-

crystalline ZrO2, metallic Ag and Cu develop at the surface and Cu is readily oxidized to

Cu2O and CuO. The outward transport of Cu and Ag is anticipated to proceed along “grain

boundaries” in the nano-crystalline ZrO2 and to be driven by the compressive stresses

induced by the volume increase associated with ZrO2 formation. In the IOZ these

compressive stresses lead to shearing in the IOZ and results in cracks inclined with respect

to the surface, which provide locations for segregation and crystallization of Cu (and Ag),

thereby “repairing” the crack. In the OOZ, the compressive stresses imposed by the

growing IOZ lead to cracks parallel to the surface, i.e. the initiation of spallation.

Segregation and crystallization of Cu and Ag, largely repairs these cracks and, in contrast

Conclusion

82

to the IOZ, Cu is subsequently oxidized as a consequence of the higher oxygen activity as

compared to the IOZ. A repetition of the spallation, Ag/Cu segregation/crystallization and

Cu oxidation leads to a stratified microstructure in the OOZ that grows into the IOZ on

prolonged oxidation.

Acknowledgements

The financial support from Villum Fonden (Grant No. 13253) is gratefully acknowledged

during this research was conducted. The authors would also like to acknowledge the

Institute of Applied Physics, Jiangxi Academy of Sciences, Nanchang, 330029, China for

providing the as-cast plate of Zr48Cu36Al8Ag8 BMG.

Manuscript I

83

References

[1] A. L. Greer and E. Ma, “Bulk Metallic Glasses : At the Cutting Edge of metals

research,” MRS Bull., vol. 32, no. August, pp. 611–619, 2007.

[2] W. H. Wang, C. Dong, and C. H. Shek, “Bulk metallic glasses,” Mater. Sci. Eng. R

Reports, vol. 44, no. 2–3, pp. 45–90, 2004.

[3] J. F. Löffler, “Bulk metallic glasses,” Intermetallics, vol. 11, no. 6, pp. 529–540,

2003.

[4] C. Suryanarayana and A. Inoue, Bulk Metallic Glasses. CRC Press, 2011.

[5] M. K. Miller and P. Liaw, Bulk Metallic Glasses: An Overview. 2007.

[6] M. F. Ashby and A. L. Greer, “Metallic glasses as structural materials,” Scr. Mater.,

vol. 54, no. 3, pp. 321–326, 2006.

[7] A. Inoue and A. Takeuchi, “Recent development and application products of bulk

glassy alloys,” Acta Mater., vol. 59, no. 6, pp. 2243–2267, 2011.

[8] J. Schroers, G. Kumar, T. M. Hodges, S. Chan, and T. R. Kyriakides, “Bulk metallic

glasses for biomedical applications,” Jom, vol. 61, no. 9, pp. 21–29, 2009.

[9] J. Eckert, J. Das, S. Pauly, and C. Duhamel, “Mechanical properties of bulk metallic

glasses and composites,” J. Mater. Res., vol. 22, no. 2, pp. 285–301, 2007.

[10] D. C. Hofmann et al., “Designing metallic glass matrix composites with high

toughness and tensile ductility,” Nature, vol. 451, no. 7182, pp. 1085–1089, 2008.

[11] G. He, J. Eckert, W. Löser, and L. Schultz, “Novel Ti-base nanostructure-dendrite

composite with enhanced plasticity,” Nat. Mater., vol. 2, no. 1, pp. 33–37, 2003.

[12] B. Sarac and J. Schroers, “Designing tensile ductility in metallic glasses,” Nat.

Commun., vol. 4, pp. 1–7, 2013.

[13] S. Haratian and M. Haddad-Sabzevar, “Thermal stability and non-isothermal

crystallization kinetics of Ti41.5Cu42.5Ni7.5Zr2.5Hf5Si1 bulk metallic glass,” J. Non.

Cryst. Solids, vol. 429, pp. 164–170, 2015.

[14] W. Zhou, J. Hou, Z. Zhong, and J. Li, “Effect of Ag content on thermal stability and

crystallization behavior of Zr-Cu-Ni-Al-Ag bulk metallic glass,” J. Non. Cryst.

Solids, vol. 411, pp. 132–136, 2015.

[15] J. Zhu et al., “Formation of Zr-based bulk metallic glass with large amount of yttrium

addition,” Intermetallics, vol. 92, no. August 2017, pp. 55–61, 2018.

[16] Y. Zhang, M. Zhou, X. Zhao, and L. Ma, “Co substituted Zr-Cu-Al-Ni metallic

glasses with enhanced glass-forming ability and high plasticity,” J. Non. Cryst.

Solids, vol. 473, no. August, pp. 120–124, 2017.

[17] T. Wada et al., “Formation and bioactivation of Zr-Al-Co bulk metallic glasses,” J.

Mater. Res., vol. 24, no. 9, pp. 2941–2948, 2009.

[18] A. Kawashima, T. Wada, K. Ohmura, G. Xie, and A. Inoue, “A Ni- and Cu-free Zr-

Conclusion

84

based bulk metallic glass with excellent resistance to stress corrosion cracking in

simulated body fluids,” Mater. Sci. Eng. A, vol. 542, pp. 140–146, 2012.

[19] L. Huang et al., “A Zr-based bulk metallic glass for future stent applications:

Materials properties, finite element modeling, and in vitro human vascular cell

response,” Acta Biomater., vol. 25, pp. 356–368, 2015.

[20] Z. Liu, K. C. Chan, and L. Liu, “Development of Ni- and Cu-Free Zr-Based Bulk

Metallic Glasses for Biomedical Applications,” Mater. Trans., vol. 52, no. 1, pp. 61–

67, 2011.

[21] J. Li, L. L. Shi, Z. D. Zhu, Q. He, H. J. Ai, and J. Xu, “Zr61Ti2Cu25Al12 metallic glass

for potential use in dental implants: Biocompatibility assessment by in vitro cellular

responses,” Mater. Sci. Eng. C, vol. 33, no. 4, pp. 2113–2121, 2013.

[22] Y. Sun et al., “Comparison of mechanical behaviors of several bulk metallic glasses

for biomedical application,” J. Non. Cryst. Solids, vol. 406, pp. 144–150, 2014.

[23] L. Liu, C. L. Qiu, C. Y. Huang, Y. Yu, H. Huang, and S. M. Zhang,

“Biocompatibility of Ni-free Zr-based bulk metallic glasses,” Intermetallics, vol. 17,

no. 4, pp. 235–240, 2009.

[24] L. Liu, C. L. Qiu, Q. Chen, K. C. Chan, and S. M. Zhang, “Deformation behavior,

corrosion resistance, and cytotoxicity of Ni-free Zr-based bulk metallic glasses,” J.

Biomed. Mater. Res. - Part A, vol. 86, no. 1, pp. 160–169, 2008.

[25] O. Baulin et al., “A Ni-, Al-, Be-free Zr-based metallic glass for biomedical

applications,” J. Non. Cryst. Solids, no. June, pp. 0–1, 2018.

[26] H. Sen Wang, W. H. Li, M. H. Wu, H. G. Chen, J. Shian-Ching Jang, and D. Y. Lin,

“The oxidation behavior of a novel Ni-free Zr-Cu-based bulk metallic glass

composite in the supercooled liquid and crystallization states,” Intermetallics, vol.

53, pp. 34–39, 2014.

[27] X. P. Nie et al., “The effect of oxidation on the corrosion resistance and mechanical

properties of a Zr-based metallic glass,” Corros. Sci., vol. 53, no. 11, pp. 3557–3565,

2011.

[28] D. Huang, L. Huang, B. Wang, V. Ji, and T. Zhang, “The relationship between t-

ZrO2 stability and the crystallization of a Zr-based bulk metallic glass during

oxidation,” Intermetallics, vol. 31, pp. 21–25, 2012.

[29] Z. Hu, L. Li, Y. Hu, J. Xing, and B. Wei, “Oxidation features of plastic deformed

Zr-based bulk metallic glass,” J. Alloys Compd., vol. 509, no. SUPPL. 1, pp. S69–

S73, 2011.

[30] W. H. Cao, J. L. Zhang, and C. H. Shek, “Oxidation behavior of Zr56Co28Al16 bulk

metallic glasses,” Corros. Sci., vol. 65, pp. 528–534, 2012.

[31] J. L. Zhang, W. H. Cao, and C. H. Shek, “Silver mushroom induced by oxidation in

Cu42Zr42Al8Ag8 metallic glasses,” J. Alloys Compd., vol. 509, no. SUPPL. 1, pp.

S219–S222, 2011.

[32] M. Zhang, D. Yao, X. Wang, and L. Deng, “Air oxidation of a Zr55Cu30Al10Ni5 bulk

metallic glass at its super cooled liquid state,” Corros. Sci., vol. 82, pp. 410–419,

Manuscript I

85

2014.

[33] K. R. Lim et al., “Oxidation resistance of the supercooled liquid in Cu50Zr50 and

Cu46Zr46Al8 metallic glasses,” J. Mater. Res., vol. 27, no. 08, pp. 1178–1186, 2012.

[34] P. Murali and U. Ramamurty, “Embrittlement of a bulk metallic glass due to sub-Tg

annealing,” Acta Mater., vol. 53, pp. 1467–1478, 2005.

[35] S. Haratian, F. Niessen, F. B. Grumsen, M. Villa, T. L. Christiansen, and M. A. J.

Somers, “Determination of residual stresses in thermochemically oxidized ZrCuAl-

based BMGs using FIB/DIC and sin2ψ methods (unpublished work).”

[36] S. Guruswamy, S. M. Park, J. P. Hirth, and R. A. Rapp, “Internal oxidation of Ag-

in alloys: Stress relief and the influence of imposed strain,” Oxid. Met., vol. 26, no.

1–2, pp. 77–100, 1986.

Conclusion

86

Manuscript II

87

6 Manuscript II

Surface hardening by gaseous oxidizing of (Zr55Cu30Al10Ni5)98Er2 bulk-

metallic glass1


Somers



Denmark

Abstract

The present investigation addresses surface hardening and the oxidation behavior of

(Zr55Cu30Al10Ni5)98Er2 bulk-metallic glass (BMG) during gaseous oxidizing below the

glass transition temperature (<Tg). The BMG was thermochemically surface engineered in

controlled gaseous atmospheres imposing either an extremely low or an extremely high

oxygen partial pressure. The hardened oxygen-containing case developing during oxidizing

the BMG was characterized with X-ray diffraction, (electron) microscopy, energy

dispersive spectroscopy and micro-hardness indentation. It was observed that oxidizing at

a high2Op resulted in the formation of an internal and an external oxidized zone; the latter

can be minimized by applying a low2Op . The inner oxide zone (IOZ) consisted mainly of

a nano-crystalline dispersion of tetragonal ZrO2, while a porous CuOx network developed

in the outer oxide zone (OOZ). The formation of copper oxide islands at the surface is

interpreted as the result of outward diffusion of metallic Cu driven by a compressive stress

gradient in the IOZ caused by volume expansion associated with ZrO2 (and Al2O3)

formation. The results demonstrate that a hardened case (the IOZ) with a maximum

hardness of ~12 GPa can be achieved. The depth distribution of oxygen is explained in

terms of a concurrent crystallization of the BMG below Tg.

1 Published work: S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, and M. A. J.

Somers, “Surface hardening by gaseous oxidizing of ( Zr55Cu30Al10Ni5)98Er2 bulk- metallic

glass,” J. Alloys Compd., vol. 800, pp. 456–461, 2019. The format of the published article

was adapted to the format of the Ph.D. thesis.

Introduction

88

6.1 Introduction

Bulk metallic glasses (BMGs), have attracted considerable scientific and technological

interest for their potential use in different functional and engineering applications,

recognizing their peculiar combination of high strength and high elastic limit [1–9]. On the

other hand, the application of BMGs is restricted by their intrinsic low fracture toughness

[10–11]. Therefore, introducing appropriate techniques for enhancing the plasticity of these

fragile materials is crucial with respect to their structural performance. Previously, efforts

were reported to promote the plasticity of metallic glasses via altering their Poisson ratio

(ν) [12–14] or synthesizing metallic glass composites with a precipitated second phases

[15–20]. Despite, the significant improvements of the fracture toughness achieved by

increasing the Poisson ratio or making amorphous matrix composites, it is of interest to

develop a practical technique to improve metallic glass plasticity without changing their

initial metallic composition and sacrificing their unique properties. Recently, several

surface treatments [21] were investigated in order to enhance the surface toughness of

metallic glasses by effectively mitigating the propagation of shear bands and crack

nucleation at the surface. These surface modification techniques, including mechanical

surface treatment [22–25], surface coating [26–28] and thermal surface treatment [29–30],

prevent shear band propagation through the introduction of compressive residual stresses

in the surface region.

Surface engineering by thermochemical treatment is widely applied to improve the

performance of alloys and metals [31]. Thermochemical treatment entails changing the

surface composition at elevated temperatures to obtain a hardened case, typically by the

incorporation of interstitial elements. Low-temperature thermochemical surface treatment

of metallic components can enhance the resistance against surface-initiated failure

mechanisms, such as fatigue, wear and corrosion.

It is noted that, heretofore, several research activities have been performed to study the

oxidation behavior of amorphous alloys in the air atmospheric conditions [32–35].

Nevertheless, thermochemical post-processing of metallic glasses in controlled oxidizing

atmospheres has so far received limited attention. In this study, the low temperature surface

hardening of (Zr55Cu30Al10Ni5)98Er2 BMG using gaseous oxidizing is investigated. It is

anticipated that the presence of Zr and Al, which both have a high affinity for oxygen and

a strongly negative Gibbs energy for oxide formation, enable the dissolution of an

Manuscript II

89

appreciable amount oxygen in the BMG, leading to a volume expansion that introduces

compressive residual stress in the surface zone.

6.2 Materials and methods

As-cast Zr-based BMG with chemical composition (Zr55Cu30Al10Ni5)98Er2 (at.%) was

provided by the Jiangxi Academy of sciences in China. The initial ingots were prepared

using vacuum arc melting of high purity (99.9 wt. %) elemental constituents in a Ti-gettered

argon atmosphere. The synthesized ingots were re-melted several times to ensure

compositional homogeneity. The homogenous ingots were subsequently cast into 2×10×60

mm3 plates using a copper mold casting method (Rapid Quench Machine System VF-

RQT50, Makabe Co. Ltd. Japan). The glass transition (Tg) and the onset crystallization

temperatures (Tx) were measured to 700 K and 740 K, respectively as determined by

differential thermal analysis in a Netzsch STA 449C thermal analyzer for a heating rate of

10 K.min-1 under a flow of argon at a flow rate of 50 cm3.min-1. Gaseous oxidizing was

carried out 10 K below the glass transition temperature in a proprietary gas mixture with

either an extremely low (2Op =10-26 bar) or an extremely high oxygen partial pressure (

2Op

=1648 bar). The oxygen partial pressures were realized chemically at a total gas pressure

of 1 bar. Samples of 3×3×1 mm3 were treated for 4 and 16 hours at 690 K (i.e. 10 K lower

than Tg) in a thermogravimetric analyzer (Netzsch STA 449 F3). The samples were

carefully polished to 1 µm diamond and then cleaned with ethanol before thermochemical

treatment. The amorphous and thermochemically oxidized samples were characterized with

X-ray diffractometry (XRD) using a Bruker D8 Discover equipped with CuKα radiation

(λ=0.15406 nm). Diffractograms were recorded in the 2θ range 25°-90° (2θ) at a step size

of 0.04° and a counting time of 8 s per step. The surface topography and the cross-section

microstructure of the thermochemically oxidized samples were examined using scanning

electron microscopy (SEM, JEOL JSM-5900) equipped with Oxford (Inca X-act) energy

dispersive spectroscopy (EDS). The microstructure of the cross-section of the surface-

treated BMG samples were investigated using a Neophot 32 (Zeiss, Jena) reflected light

microscope. A JEOL 3000F transmission electron microscope (TEM) coupled with EDS

detector was employed for supplementary characterization of the various zones formed in

the oxidized BMG specimen treated at 690 K for 16 hr in a high 2Op atmosphere. An

electron transparent TEM foil was prepared using a FEI Helios Nanolab with focused ion


90

beam (FIB) milling using Ga+ ions and an Omniprobe micro-manipulator. Initial rough FIB

milling was carried out with a 30 kV ion beam acceleration voltage and 20 nA ion current.

It should be noticed that the region of interest on the sample surface was protected from the

excessive ion damage with the Pt-deposited layer. Thereafter, the lamella was lifted out

with the Omniprobe and then mounted to a TEM Cu grid. The thinning process was

performed using an ion beam current of 90 pA-0.9 nA on both sides of the extracted lamella.

Finally, a low acceleration voltage (2 kV) cleaning was done with a 24 pA ion beam current

to remove the residual Ga+ ion damage on the lamella. To evaluate the hardened case,

Vickers micro-hardness indentations (standard E 384-17) were determined on the cross-

section of the thermochemically oxidized samples, using a FutureTech FM-700 Vickers

indenter with 5 gf load and a dwell time of 10 s.


X-ray diffraction of as-cast and thermochemically oxidized BMG

The X-ray diffractograms of the as-cast (Zr55Cu30Al10Ni5)98Er2 BMG and the specimens

thermochemically oxidized in low and high oxygen partial pressure are shown in Fig. 6.1.

The as-cast material has a broad peak at 2θ≈38º, indicating that it is amorphous without

detectable crystalline phases. The XRD results exhibit that after isothermal heating at 690

K for 4 hr in a low oxygen partial pressure atmosphere, the amorphous phase devitrifies

and forms crystalline Zr-Cu, Zr-Cu-Al and Zr-Ni intermetallic phases, while oxidation

results in the formation of tetragonal-ZrO2 (t-ZrO2), which is ascribed to the high affinity

of Zr for O. Prolonging the treatment to 16 hr, the intensity of t-ZrO2 is amplified and only

a small fraction of other crystalline phases can be detected. Evidently, the intermetallic

compounds are located beyond the information depth of the applied X-rays. Further, the

oxidation at a high chemical partial pressure of oxygen under identical isothermal heating

condition leads to the formation of t-ZrO2 and three types of copper oxides: CuO, Cu2O,

and Cu4O3. Interestingly, in addition to the copper oxides formed during the

thermochemical treatment, also diffraction peaks of pure copper are distinguished in the

diffractograms. For the specimens oxidized at high2Op , no intermetallic phases were found

in the X-ray diffractogram, implying that the oxide layer is relatively thick as compared to

the information depth of the applied X-radiation. From the intensity reduction of the main

Manuscript II

91

t-ZrO2 peak (2θ≈30º) on prolonged oxidizing at high 2Op atmosphere it is concluded that

the amount of CuOx at the surface increases with oxidizing time.

Fig. 6.1. The X-ray diffractograms of the as-cast and thermochemically oxidized (Zr55Cu30Al10Ni5)98Er2 (at.

%) BMG.

Cross-sectional hardness measurement

Microhardness-depth profiles obtained after oxidizing at 690 K for 4 hr and 16 hr in the

thermogravimetric analyzer are given in Fig. 6.2. The effective case depths after oxidizing

for 4 and 16 hr at low 2Op are ~2 µm and ~10 µm, respectively, while at high

2Op case

depths of ~18 µm and ~34 µm were reached (see dotted vertical lines in Fig. 6.2). The

hardness of the treated BMG reaches approximately ~12 GPa close to the surface, and

decreases gradually until it reaches a steep drop to the hardness value of the un-oxidized

BMG at the case-to-core transition. Evidently, changing the oxygen partial pressure has a

large influence on the thickness of the hardened case depth.


92

Fig. 6.2. Microhardness profiles as a function of depth for (Zr55Cu30Al10Ni5)98Er2 BMG thermochemically

oxidized at 690 K for 4 and 16 hr in (a) low (b) high oxygen partial pressure atmospheres (pO2).

Microstructure characterization of the oxidized zones

Microscopical investigations were carried out on the BMG that was thermochemically

oxidized at 690 K for 16 hr in a high 2Op atmosphere. A secondary electron micrograph of

the surface topography is presented in Fig. 6.3-a. A gray porous oxide network appears to

cover the surface. The chemical composition of the oxide network was probed with EDS

point-to-point analyses and showed copper and oxygen, while other elements were virtually

absent, indicating that the porous oxide network is formed by copper oxides. A light-optical

image of a cross-section over the hardened case is provided in Fig. 6.3-b and shows the

presence of island-like (Cu/CuO) particles lying on the top of the sample surface. Also, a

Cu-colored line directed perpendicularly to the surface is identified (see white arrow). This

feature is interpreted as a Cu-filled crack (see discussion). For further cross-sectional

inspection of the oxidized (Zr55Cu30Al10Ni5)98Er2 BMG Fig. 6.3-c shows a back-scatter

electron (BSE) micrograph and Fig. 6.3-d EDS maps of the area marked in Fig. 6.3-c. In

addition, Fig. 6.3-e shows composition-depth profiles determined with EDS. In the

quantification of these profiles only the metallic components were considered, while the

oxygen profile represents oxygen intensity relative to the maximum value. In Fig. 6.3-c and

e, the case is subdivided in an inner oxide zone (IOZ) and an outer oxide zone (OOZ), based

on the fraction of Cu revealed by EDS. The OOZ consists of island-like particles lying on

the surface and of the top surface of the sample, appearing dark in Fig. 6.3-c.

Manuscript II

93

Fig. 6.3. (a) Secondary electron (SE) micrograph of surface morphology of (Zr55Cu30Al10Ni5)98Er2 BMG

thermochemically oxidized at 690 K for 16 hr in high pO2 atmosphere; (b) Optical image of the cross-section

view demonstrating a self-healed microcrack with metallic Cu; (c) Cross-sectional BSE micrograph of the

oxide zones formed on (Zr55Cu30Al10Ni5)98Er2 BMG after oxidizing treatment at 690 K for 16 hr (d) EDS

mapping of the selected area in c; (e) atomic fraction of metallic components (left vertical scale) and oxygen

intensity profile (right vertical scale).

Additionally, the IOZ is subdivided in three distinct zones denoted as I, II and III. The EDS

results (Figs. 6.3-d and e) show that the OOZ only contains copper and some oxygen. In

the entire IOZ, the oxygen content decreases gradually while zirconium and aluminum are

more or less homogeneously distributed. It is anticipated that the trend of increasing Zr and

Al contents towards the surface in the IOZ, is a consequence of changes in density caused

by internal oxidation and omitting O in the quantification of the EDS intensities of the

metallic components. Zones I and II in the IOZ are closest to the surface and are depleted

in copper. A sharp discontinuity in the Cu and Ni contents marks the transition between

zones II and III in the EDS profile (marked by white arrow in Ni map). The oxygen content

is highest in zone I and slightly lower (and constant) in zone II. In zone III the oxygen

content shows a maximum at the location where the Cu content goes through a minimum

and thereafter decreases with distance to the surface. Along with the decrease in oxygen

content the BSE micrograph in Fig. 6.3-c becomes brighter, indicating an increase in atomic

Discussion

94

density. The gradient in oxygen content is interpreted as a reduction of the amount of Zr

(and Al) available for internal oxide formation, because of the development of intermetallic

compounds of Zr and Al prior to the arrival of oxygen. Although oxide formation

eventually will prevail thermodynamically, the nucleation of ZrO2 is likely to be delayed,

because i) the driving force for oxide formation is reduced if Zr is present in an

intermetallic compound and ii) the reduction in free volume associated with devitrification

and crystallization enhances strain energy evoked by internal oxidation. The transition

from IOZ to oxygen-free BMG is very sharp, consistent with a sharp case-core transition

in the hardness profile in Fig. 6.2-b.

Another feature observed in the cross-section BSE micrograph in Fig. 6.3-c is the presence

of cracks perpendicular to the surface in zone I (dark lines marked by white arrows). Such

cracks hint at a tensile stress imposed onto zone I.

Cross-sectional TEM bright-field images and their corresponding selected area electron

diffraction (SAED) patterns are shown in Fig. 6.4 for the OOZ and in zone II of the IOZ.

Closest to the surface the SAED patterns show the presence of monoclinic CuO and

tetragonal Cu4O3, while zone I in the IOZ contains nano-crystalline t-ZrO2.

6.4 Discussion

The results obtained on oxidizing (Zr55Cu30Al10Ni5)98Er2 BMG at a temperature below Tg

show that the incorporation of oxygen leads to the development of an IOZ consisting of

nano-crystalline t-ZrO2 within the BMG and an OOZ consisting of Cu-based oxides.

Although the oxidizing temperature was chosen 10 K below Tg, the development of

crystalline intermetallic phases was observed for the sample oxidized at a low 2Op for 4h,

where the thinnest IOZ has developed. It is likely that such crystallization also occurred for

the other oxidizing condition, but it remained unobserved because of the limited

information depth of the applied X-radiation. Nevertheless, under the present oxidizing

conditions the developing IOZ is the result of a competition between the crystallization of

intermetallic compounds and internal oxidation of strong oxide forming alloying elements.

Manuscript II

95

Fig. 6.4. Cross-sectional TEM bright-field image of thermochemically oxidized (Zr55Cu30Al10Ni5)98Er2 BMG

after applying the treatment at 690 K for 16 hr and their typical corresponding SAED patterns which have

been obtained from different positions: (a) position 1 (b) position 2 (c) position 3.

The strongest oxide formers in the investigated BMG are Zr and Al. The Gibbs energy for

oxide formation (per mol O2) is slightly more negative for ZrO2 than for Al2O3. Hence, for

the present BMG, which contains ~55 at.% Zr and ~10 at.% Al, ZrO2 is the

thermodynamically most stable oxide, which could explain why it is the only oxide present

after oxidation at a low 2Op for 4h (Fig. 6.1). As no indications for the development of

Discussion

96

Al2O3 were found, not even after prolonged oxidation or at higher 2Op , it is concluded that

Al2O3 is either amorphous or it is dissolved in ZrO2 [36].

The case depth of the IOZ was observed to depend strongly on the 2Op in the oxidizing

gas (Fig. 6.2 and section 6.3.2). For the low2Op , the hardened case depth is increased about

4 times on prolonging the oxidizing treatment by a factor 4; for the high2Op , the hardened

case depth is about doubled when the duration is quadrupled. These observations suggest

that for the oxidizing conditions at low2Op , growth of the case depth (the IOZ) is not

diffusion controlled, but rather controlled by the surface reaction, i.e. the transfer of oxygen

from the gas to the solid. On the other hand, the observations for the high 2Op are

consistent with diffusion-controlled growth of the IOZ. The diffusing species that governs

the thickness of the IOZ for the high 2Op is most likely atomic oxygen.

The development of an OOZ consisting of Cu-based oxides and a redistribution of Cu

in the IOZ (especially zones I and II) is explained as follows. The development of nano-

crystalline t-ZrO2 in the IOZ leads to a volume expansion. This volume expansion is

accommodated by compressive stress in the IOZ, which leads to a higher chemical potential

for the components in IOZ than in the unaffected regions, i.e. the surface and the interior

of the BMG. Consequently, a driving force for diffusion out of the IOZ is established and

those components which are not bound to oxides, mainly Cu and Ni, will diffuse out of the

IOZ. For oxidizing at high 2Op the Cu arriving at the surface will oxidize to Cu-oxide and

develop the porous network that forms the OOZ, while Cu does not oxidize at low2Op .

Also the observation of metallic Cu in microcracks can be explained by the outward

diffusion of Cu towards a free (crack) surface (Fig. 6.3-b). If no oxidation occurs of Cu it

implies that the local 2Op is not sufficiently high to stabilize Cu-oxide. Thus, microcracks

can in principle be “repaired” by this self-healing mechanism. This phenomenon has also

been recently reported for air-oxidation of the noble-metal containing Zr-based BMG by

the present authors [37].

The development of micro-cracks perpendicular to the surface was only observed in

zone I of the IOZ, and indicates that tensile stresses have been present in this region. In

Manuscript II

97

order to understand how these tensile stresses were introduced in zone I of the IOZ it is

important to explain the subdivision of the IOZ in zones I, II and III. The subdivision is

most likely associated with the competition between internal oxidation and the nucleation

of intermetallic compounds and devitrification in the core. Clearly, the volume changes in

the various zones within the IOZ will be different. Apparently, zone I has experienced a

volume expansion imposed by the underlying zones II and III while they expanded on

internal oxidation. This is suggested to be caused as follows. Presuming that no

intermetallic compounds formed in zone I prior to internal oxidation of Zr (and Al), the

devitrification deeper in the material will lead to additional compressive stress imposed

onto zone I, resulting in an additional flux of Cu out of this region to comply with these

stresses and reduce them. Then, on subsequent formation of ZrO2 deeper in the (partly)

devitrified and crystallized BMG, the occurring volume expansion leads to imposing tensile

straining of zone I, causing crack initiation perpendicular to the surface.

Obviously, the hardness profiles in Fig. 6.2 correlate with the oxygen content in the

IOZ (Fig. 6.3-e). The highest oxygen content corresponds to the highest hardness in zone

I. In zone II the hardness is constant at about 1060 HV, while the hardness decreases

gradually in zone III along with the reduction in oxygen content, to fall abruptly at the

transition from IOZ to oxygen-free alloy.

6.5 Conclusion

The effect of gaseous oxidizing as a thermochemical treatment of

(Zr55Cu30Al10Ni5)98Er2 BMG below the glass transition temperature (Tg) was investigated

for different treatment durations in low and high oxygen partial pressure (2Op ). It was

demonstrated that for oxidizing at a low oxygen pressure internal oxidation of Zr to

tetragonal ZrO2 occurs in competition with crystallization of the BMG, despite an oxidizing

temperature below Tg. The growth kinetics of the inner oxide zone (IOZ) for low 2Op

appears to be linear, suggesting that the oxygen transfer to the BMG is rate determining.

For high 2Op two different oxide zones formed at the surface: an outer oxide zone (OOZ)

consisting of porous CuOx islands and an IOZ which is subdivided in three subzones. The

presence of copper oxides in the OOZ is attributed to the outward diffusion of copper from

the inner oxide zone as a result of compressive growth stresses induced by the formation

Conclusion

98

of ZrO2. The subdivision of the IOZ is suggested to be the outcome of the competition

between internal oxidation of Zr (and Al) and devitrification and crystallization of the

BMG. Parabolic growth kinetics of the IOZ appears to apply, indicating (oxygen) diffusion

controlled growth.

The dissolution of oxygen into the MG substrate effectively enhances the surface

hardness by forming the ZrO2 in the IOZ. The hardness in the oxygen-containing case

increased significantly from approximately ~ 600 HV in the core of un-oxidized BMG to

~1200 HV.

Acknowledgements

Villum Fonden is gratefully acknowledged for financial support under Grant No. 13253.

The authors acknowledge the Institute of Applied Physics, Jiangxi Academy of Sciences,

Nanchang in China for providing the starting material.

Manuscript II

99

References


vol. 54, no. 3, pp. 321–326, 2006.

[2] M. M. Trexler and N. N. Thadhani, “Mechanical properties of bulk metallic glasses,”

Prog. Mater. Sci., vol. 55, no. 8, pp. 759–839, 2010.

[3] A. R. Yavari, J. J. Lewandowski, and J. Eckert, “Mechanical Properties of Bulk

Metallic Glasses,” vol. 32, no. August, pp. 635–638, 2007.



[5] M. Telford, “The case for bulk metallic glass,” Mater. Today, vol. 7, no. 3, pp. 36–

43, 2004.





[8] W. Zhang, A. Inoue, and X. M. Wang, “Developments and Applications of Bulk

Metallic Glasses,” Rev. Adv. Mater. Sci., vol. 18, pp. 1–9, 2008.

[9] H. F. Li and Y. F. Zheng, “Recent advances in bulk metallic glasses for biomedical

applications,” Acta Biomater., vol. 36, pp. 1–20, 2016.

[10] Z. Q. Liu and Z. F. Zhang, “Strengthening and toughening metallic glasses: The

elastic perspectives and opportunities,” J. Appl. Phys., vol. 115, no. 16, pp. 0–11,

2014.

[11] J. Xu, U. Ramamurty, and E. Ma, “The fracture toughness of bulk metallic glasses,”

Jom, vol. 62, no. 4. pp. 10–18, 2010.

[12] S. V. Madge, D. V. Louzguine-Luzgin, J. J. Lewandowski, and A. L. Greer,

“Toughness, extrinsic effects and Poisson’s ratio of bulk metallic glasses,” Acta

Mater., vol. 60, no. 12, pp. 4800–4809, 2012.

[13] S. J. Poon, A. Zhu, and G. J. Shiflet, “Poisson’s Ratio and Intrinsic Plasticity of

Metallic Glasses,” Appl. Phys. Lett., vol. 92, no. 26, pp. 2006–2009, 2008.

[14] G. N. Greaves, A. L. Greer, R. S. Lakes, and T. Rouxel, “Poisson’s ratio and modern

materials,” Nat. Mater., vol. 10, no. 11, pp. 823–837, 2011.

Conclusion

100

[15] D. C. Hofmann, “Bulk Metallic Glasses and Their Composites: A Brief History of

Diverging Fields,” J. Mater., vol. 2013, p. e517904, 2013.



[17] J. Z. Jiang, D. Hofmann, D. J. Jarvis, and H. J. Fecht, “Low-Density High-Strength

Bulk Metallic Glasses and Their Composites: A Review,” Adv. Eng. Mater., vol. 17,

no. 6, pp. 761–780, 2015.

[18] G. He, J. Eckert, W. Löser, and L. Schultz, “Novel Ti-base nanostructure-dendrite

composite with enhanced plasticity,” Nat. Mater., vol. 2, no. 1, pp. 33–37, 2003.

[19] J. Eckert et al., “High strength ductile Cu-base metallic glass,” Intermetallics, vol.

14, no. 8–9, pp. 876–881, 2006.

[20] J. Eckert, J. Das, S. Pauly, and C. Duhamel, “Mechanical properties of bulk metallic

glasses and composites,” J. Mater. Res., vol. 22, no. 2, pp. 285–301, 2007.

[21] T. G. Nieh, Y. Yang, J. Lu, and C. T. Liu, “Effect of surface modifications on shear

banding and plasticity in metallic glasses: An overview,” Prog. Nat. Sci. Mater. Int.,

vol. 22, no. 5, pp. 355–363, 2012.

[22] Y. Zhang, W. H. Wang, and A. L. Greer, “Making metallic glasses plastic by control

of residual stress,” vol. 5, no. November, pp. 857–860, 2006.

[23] J. Fu, Y. Zhu, C. Zheng, R. Liu, and Z. Ji, “Effect of laser shock peening on the

compressive deformation and plastic behavior of Zr-based bulk metallic glass,” Opt.

Lasers Eng., vol. 86, pp. 53–61, 2016.

[24] C. Ma et al., “Increasing fracture strength in bulk metallic glasses using ultrasonic

nanocrystal surface modification,” vol. 718, pp. 246–253, 2017.

[25] F. O. Méar, B. Doisneau, A. R. Yavari, and A. L. Greer, “Structural effects of shot-

peening in bulk metallic glasses,” vol. 483, pp. 256–259, 2009.

[26] W. Chen, K. C. Chan, P. Yu, and G. Wang, “Encapsulated Zr-based bulk metallic

glass with large plasticity,” Mater. Sci. Eng. A, vol. 528, no. 6, pp. 2988–2994, 2011.

[27] W. Chen, K. C. Chan, S. H. Chen, S. F. Guo, W. H. Li, and G. Wang, “Plasticity

enhancement of a Zr-based bulk metallic glass by an electroplated Cu / Ni bilayered

coating,” vol. 552, pp. 199–203, 2012.

[28] L. W. Ren et al., “Enhancement of plasticity in Zr-based bulk metallic glasses

electroplated with copper coatings,” Intermetallics, vol. 57, pp. 121–126, 2015.

Manuscript II

101

[29] J. Fan, A. Chen, J. Wang, J. Shen, and J. Lu, “Improved plasticity and fracture

toughness in metallic glasses via surface crystallization,” Intermetallics, vol. 19, no.

10, pp. 1420–1427, 2011.

[30] G. Wu et al., “Induced multiple heterogeneities and related plastic improvement by

laser surface treatment in CuZr-based bulk metallic glass,” Intermetallics, vol. 24,

pp. 50–55, 2012.

[31] E. J. Mittemeijer and M. A. J. Somers, Thermochemical Surface Engineering of

Steels. Woodhead Publishing Limited, 2015.

[32] Z. Xu, Y. Xu, A. Zhang, J. Wang, and Z. Wang, “Oxidation of amorphous alloys,”

J. Mater. Sci. Technol., vol. 34, no. 11, pp. 1977–2005, 2018.

[33] Y. Xu et al., “Natural oxidation of amorphous Cux Zr1-x alloys,” Appl. Surf. Sci., vol.

457, no. June, pp. 396–402, 2018.

[34] W. Kai, F. P. Cheng, Y. T. Chen, R. T. Huang, H. H. Huang, and W. Zhang, “Air-

oxidation of an amorphous Fe70C10P10B5Mo5 alloy at 400–500 °C,” J. Alloys

Compd., vol. 763, pp. 209–215, 2018.



[36] S. Moreau, M. Gervais, and A. Douy, “Formation of metastable solid solution in the

ZrO2 rich part of the system ZrO2-Al2O3.” pp. 625–631, 1997.

[37] S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, and M. A. J. Somers, “Self-

repair by stress-induced diffusion of noble elements during oxidation of

Zr48Cu36Al8Ag8 bulk metallic glass,” Scr. Mater., vol. 164, pp. 126–129, 2019.

Conclusion

102

Manuscript III

103

7 Manuscript III

Stress, stress relaxation and self-healing in (Zr55Cu30Al10Ni5)98Er2 bulk metallic glass

during air-oxidation1


Somers



Denmark

Abstract

The relation between the microstructure evolution and the state of the stress in the surface

region of (Zr55Cu30Al10Ni5)98Er2 BMG during air-oxidation was investigated. The

oxidation was followed in-situ with X-ray diffraction (XRD) and ex-situ with various

electron microscopy techniques. In-situ XRD reveals the evolution of crystalline oxide

phases as well as the development of compressive residual stresses in the surface region of

the oxidized BMG. Post-oxidation microscopical investigations show that the dissolution

of oxygen into the surface zone of the BMG results in the formation of two oxidation

regions, consisting of an outer oxide zone (OOZ), comprised of a CuOx network on the

surface, and an internal oxidation zone (IOZ), consisting mainly of nano-crystalline

tetragonal ZrO2 showing, Cu/CuOx as well as micro-cracks. It is suggested that the

compressive residual stresses developed because of a growing internal oxidation zone ~1.4

GPa are (partly) relaxed by a redistribution of Cu. The results demonstrate that internal

micro-crack surfaces are effectively sealed by Cu segregation, a surprising self-healing

effect.

1 Unpublished work at the date of Ph.D. thesis submission. The final article may deviate

from the present manuscript.

Introduction

104

7.1 Introduction

Bulk metallic glasses (BMGs) are kinetically frozen metastable amorphous solids without

the long-range order among the atoms that characterizes crystalline materials [1–4]. The

non-crystalline nature of BMGs makes them unique materials with notable mechanical and

physical properties in order to be exploited in current and future technologies [5–10]. Up

to now, extensive research activities have led to the development of several multi-

component glassy alloying systems including Pd-, Fe-, Zr-, Ti- and Cu- based BMGs [11–

12]. Among them, ZrCuAl-based BMGs are of major interest and most promising systems

due to a high glass-forming ability (GFA), thermal stability, and good mechanical

properties [13–14]. At the same time, the presence of strong oxide-forming elements like

Zr and Al in the alloying system makes them thermodynamically favorable for oxidation

in oxygen-containing environments [15–16]. Therefore, it is scientifically interesting to

clarify the metallic glass surface response when it is exposed to an oxygen-providing

atmosphere at elevated temperature. The scope of some previous works has been limited to

studying the oxidation kinetics and the oxidation-induced crystallization of such interface-

free materials [17–19]. To the best of authors’ knowledge, almost all previous

investigations were performed ex-situ (post-mortem), studying the effect of oxidation

parameters, i.e. temperature, time, and the oxygen partial pressure on the formation of

oxidic phases, after cooling to room temperature. So far, the oxidation behavior of binary

[20], ternary [21], and multicomponent [22] Zr-based BMGs has been investigated. In-

depth understanding of the oxidation mechanism and the interaction between the evolution

of internal stresses and the growth of the surface-oxidized scale using in-situ (direct

monitoring) techniques has received practically no attention.

Recently, long-term oxidation experiments at temperatures below the glass transition

temperature (Tg) in different oxidizing environments were performed on ZrCuAl-based

BMGs and aimed at determining how the dissolution of oxygen in the glassy metal leads

to the development of one or more oxidized zones [23–24]. These studies were not only

conducted to investigate the microstructure evolution during oxidation, but also targeted at

enhancing the BMG’s surface hardness and at establishing compressive residual stresses in

the surface region, hence improving the material’s resistance against crack initiation [23].

It has been suggested that the observed microstructures develop as a consequence of

oxidation and, simultaneously, of the development of residual stresses and their (partial)

Manuscript III

105

relaxation. Currently, no information is available concerning how stresses develop, how

they are distributed in the various crystalline oxide phases, and whether actual stress

relaxation does occur. Therefore, the present investigation aims at establishing a better

understanding of the oxidation mechanism of a multicomponent ZrCuAl-based BMG,

where the various elements possess different affinities for oxygen. For this purpose, the

development of the oxide phases and the stress evolution during long-term sub-Tg air-

oxidation of (Zr55Cu30Al10Ni5)98Er2 were studied using in-situ XRD. The microstructural

characterization was applied post-mortem by means of ex-situ (electron) microscopy

techniques.

7.2 Materials and Methods

Sample preparation and oxidation experiments

The ZrCuAl-based BMG used for the investigation was (Zr55Cu30Al10Ni5)98Er2. The

synthesis procedure was presented elsewhere [23].

Isochronal (constant heating rate) differential thermal analysis (DTA) was applied to

determine the thermal properties of the studied BMG using a Netzsch STA 449 C

differential thermal analyzer at 10 K/min under a flow of pure argon at a flow rate of 50

cm3/min. X-ray diffraction (XRD) was applied to confirm the amorphous structure of the

BMG using a Bruker D8 Discover X-ray diffractometer with parallel beam geometry and

a Cu Kα radiation (λ=0.15418 nm). The applied 2θ-range was 20°-90° at the step size (∆θ)

and counting time per step 0.04° and 8 s, respectively.

Two air-oxidation experiments were carried out. In the first oxidation experiment, the

evolution of the phases developing during oxidation was followed in-situ in a Bruker D8

Discover diffractometer, which this time was equipped with Cr Kα radiation (λ=0.229 nm).

The experiment, performed on a 3×1×0.5 mm3 BMG specimen, comprised three main

stages: i) isochronal heating to 600 K in 20 K intervals, ii) isothermal holding for 60 hr and

iii) cooling to room temperature at 50 K intervals. After each temperature step, XRD was

applied in the 2θ range 42°-75°. This angular range was selected to include all possible

characteristic oxide reflections that form during air-oxidation of the BMG. A step size of

0.06° (∆2 and a counting time per step of 3 s were used for recording the diffractograms.

The temperature accuracy during in-situ oxidation experiments was ±10 K. Additional

Materials and Methods

106

details on the experimental setup can be found in Ref. [24]. Moreover, a detailed XRD scan

of the oxidized BMG sample was acquired post-mortem after in-situ oxidation experiment

using Cu Kα radiation. The acquisition was performed in a scattering angle, 2θ, ranging

20°-90° with a step size of 0.04° and counting time of 8 s per step.

In the second experiment, in situ XRD was applied to follow the evolution of the state of

stress during oxidation by applying the sin2ψ method. The lattice strain was determined in

symmetrical beam incident geometry, using a point focus configuration (collimator size: 1

mm) and a PolyCap for near-parallel beam geometry. For this analysis, the diffractometer

was equipped with Cu-Kα (λ=0.15418 nm). A ~ 20×5×1 mm3 BMG specimen polished to

1 µm diamond was fixed on a PtRh-heating band using two Alumina rods. A temperature

controller was plugged in the heating stage and an additional thermocouple was mounted

on the sample to monitor the BMG surface temperature. The thermal cycle was kept

comparable to the one applied in experiment 1 but for the heating step, which was, overall,

much slower. Specifically, the sample was firstly heated to 370 K and then to 600 K in 20

K intervals2; at each step, a series of diffractograms were acquired at 5 ψ (side-inclination)

angles, corresponding to sin2ψ= (0, 0.15, 0.30, 0.45, 0.60), by tilting the sample over the

normal to the direction of the beam. The measurements were performed in the 2θ range

26°-47°. This 2θ range was chosen to include various reflections of the oxide phases, i.e.

crystallographic planes of {011} tetragonal ZrO2, {111} monoclinic CuO, and {111} cubic

Cu2O. The step size of 0.04 (∆θ°) and the counting time per step 3 s were selected such that

the total acquisition time per stress measurement (5 diffractograms) was about 2.2 h. After

reaching 600 K, oxidation of the BMG was followed isothermally for 60 h. The state of

stress was probed intermittently by applying a series of sin2ψ measurements under the same

acquisition conditions as applied during heating (i.e. one per ~ 2.2 h). Finally, the

investigation was continued during cooling to 300 K, in steps of 50 K. To obtain the peak

positions, X-ray diffraction line profiles were fitted with Pseudo-Voigt functions using the

TOPAS P2-1 software (Bruker AXS).

2 No (micro-) structural changes in the investigated amorphous alloying system were

expected to take place below 370 K. Consequently, the initial temperature for the sin2ψ

measurement was set to 370 K.

Manuscript III

107

Microscopical characterization of the oxidation zones

The investigation was performed on the sample subjected to experiment 1. The cross-

sectional microstructure was characterized using a Neophot 32 (Zeiss, Jena) light optical

microscope (LOM) and a Supra 35 scanning electron microscope (SEM, Zeiss) coupled

with a Thermo NORAN system 6 for energy dispersive X-ray spectroscopy (EDS). The

surface (top-down view) of the sample was investigated in an FEI Helios Nanolab 600 field

emission gun scanning electron microscope (FEG-SEM), equipped with a focused ion

beam (FIB). The oxidation zone that has developed during the in-situ XRD experiment was

further characterized in a JEOL 3000F transmission electron microscope (TEM) fitted with

an Oxford Instruments EDS detector. A TEM lamella was extracted, covering the first~ 5

μm below the surface of the oxidized sample, and was prepared for electron transparency

using the available Ga+ ion source in FEI Helios Nanolab 600 microscope. A detailed TEM

lamella preparation procedure used in this study can be found in Ref. [23].


Characterization of as-cast and as-oxidized BMG samples

The DTA thermogram of (Zr50Cu30Al10Ni5)98Er2 recorded during continuous heating in a

flow of purified Ar (Fig. 7.1-a) shows the onset of an exothermic crystallization peak at

748 K (onset Tx). Considering the change in the slope of the thermogram before

crystallization, Tg was assessed as 700 K. Both Tg and onset Tx temperatures are indicated

by arrows in Fig. 7.1-a. The X-ray diffractograms of the as-cast and the in-situ air-oxidized

(Zr55Cu30Al10Ni5)98Er2 BMG (see further in section 7.3.2) are given in Fig. 7.1-b. The

presence of two amorphous humps at 2θ ≈ 38º and 65º indicates the lack of crystallinity in

the investigated BMG. The XRD result shows that the atmospheric oxidation results in the

formation of tetragonal ZrO2, CuOx i.e. monoclinic CuO and cubic Cu2O phases and a tiny

fraction of NiO and metallic Cu in the surface region of the BMG. No indications of the

formation of crystalline intermetallic phases were observed within the information depth

range probed.


108

Fig. 7.1. (a) DTA thermogram of the (Zr55Cu30Al10Ni5)98Er2 BMG acquired during continuous

heating at a rate of 10 K/min using pure Ar as a protective gas. (b) X-ray diffractograms of the as-

cast and in-situ air-oxidized BMG.

In-situ XRD air-oxidation

In order to investigate the evolving phases during long-term oxidation, in-situ X-ray

diffraction was employed. A 2D representation of the XRD results is given in Fig. 7.2. The

representation shows the evolution of X-ray diffractograms with time, including heating to

600 K, isothermal holding at 600 K for 60 hr, and isochronal cooling. The diffracted

intensity is represented by the color scale. In the early stages of heating, the sample is

amorphous as reflected by the broad peak marked by the dotted line. The intensity of the

amorphous hump decreases with increasing temperature while crystalline oxide phases

form in the surface region of the BMG. On heating, tetragonal ZrO2 forms as an initial

oxide at ~470 K, reflecting the high affinity of Zr for O. On continued oxidation, the

intensity of ZrO2 is reduced while the intensity of Cu2O and CuO diffraction peaks

increases pronouncedly. This is explained from CuOx phases forming at the surface atop

ZrO2. Interestingly, a diffraction line of metallic Cu becomes visible, but this peak

disappears shortly and is replaced by two types of Cu-oxides. On continued oxidation the

intensities of the diffraction lines for CuO become higher than for the Cu2O diffraction

peaks, implying that CuO is the most stable and dominating phase of the Cu-oxides and

most likely is closest to the surface. The results are in excellent agreement with a previous

study on in-situ long-term air-oxidation of a Zr48Cu36Al8Ag8 BMG, where it was shown

Manuscript III

109

that an outer oxide zone (OOZ) develops atop an inner oxidization zone (IOZ); the latter

mainly contains nano-crystalline ZrO2 [24].

On cooling, the diffraction peaks of ZrO2 and CuO shift to higher 2θ values because of

thermal contraction, while the position of the Cu2O peaks shifts slightly to the lower 2

angles. This anomalous behavior was also observed during in-situ XRD air-oxidation of

Zr48Cu36Al8Ag8 BMG [24] and was rationalized by a negative thermal expansion

coefficient (NTE) of Cu2O, which has been reported for low temperatures (9-240 K)3.

Alternatively, this shift to lower 2 is a consequence of thermal shrink counteracted by

(additional) compressive stresses.

Fig. 7.2. The 2D plot of in-situ XRD oxidation results of (Zr55Cu30Al10Ni5)98Er2 BMG during heating,

isothermal holding at 600 K for 60 hr and cooling in the air at atmospheric pressure. The square root of

diffracted intensities is presented by the color-scale on the right.

Microscopical characterization of the oxidation zones

Light-optical and scanning electron microscopy

The microstructure of the sample after oxidation was characterized by various

microscopical techniques. Secondary electron (SE) micrographs of the oxidized surface are

presented in Fig. 7.3-a. The surface is covered by a network of oxide ridges. In-between

the ridges globular oxide regions are observed. On top of the oxide, fine oxide whiskers

have developed. Selective growth of CuO nanowires (or whiskers) after air-oxidation of a

Cu60Zr30Ti10 BMG was earlier reported in Ref. [25]; it is therefore anticipated that also the

3 This temperature range is beyond that investigated here.


110

whiskers in Fig. 7.3-a are CuO. At several locations cracks, marked by white arrows in Fig.

7.3-a (enlargement on the right), were observed in the oxidized layer. The chemical

composition of the surface oxides was examined using point-to-point (EDS) analyses and

assessed to be enriched in copper and oxygen; other metallic elements from the BMG were

not detected with EDS applied to the surface oxides. The EDS results are in line with the

in-situ XRD results, where Cu-based oxides were detected at the surface (see section 7.3.2).

The development of a surface-network of CuOx for the same Er-containing BMG was also

observed after oxidation at an extremely high oxygen partial pressure (pO2) [23].

Micrographs obtained with cross-sectional light optical microcopy at three different

locations are presented in Fig. 7.3-b. Two distinct oxidation zones are observed and

designated as internal oxidation zone (IOZ) and outer oxide zone (OOZ), respectively. The

OOZ consists of irregularly distributed mushroom-like Cu/CuOx protruding from the IOZ.

Within the IOZ, a relatively light zone, closest to the surface, and a relatively dark zone,

are distinguished. This contrast difference is further corroborated by back-scattered

electron contrast and the elemental distribution maps, leading to the division of the IOZ

into two sub-zones, and specified as I and II in Fig. 7.3 b-c, interspersed by the narrow

transition zone. The entire IOZ appears depleted in copper. Zone I, closest to the surface,

appears darkest in the BSE micrograph (Fig. 7.3-c) and has the lowest Cu intensity in the

EDS map (Fig. 7.3-d). Zone II, marked by the black arrow in the EDS map, appears

depleted in Ni distribution. Micro-cracks developed perpendicularly to the sample surface

in the near-surface Cu-depleted zone (Fig. 7.3 b-c). Some of the micro-cracks appear to

have nucleated from the transition of zone I to zone II in IOZ, while some smaller appear

to have initiated at the surface (see discussion). Copper-containing micro-cracks are

marked by white and black arrows in the micrographs in Fig. 7.3-b and d indicate that

segregation and crystallization of Cu have occurred at the crack surface. At several

locations, Cu is present in metallic form, as indicated by the natural red color of Cu in the

light-optical micrographs in Fig. 7.3-b, particularly in the micrograph of the sample corner

on the right. It is anticipated that the volume expansion and associated local stress

concentration at a sharp corner of the sample, where oxygen diffuses in through two

surfaces, leads to the formation of a wedge-shaped crack (geometrical effect). Cu-

enrichment in the OOZ and in micro-cracks in the IOZ is confirmed by EDS analysis (see

Fig. 7.3-d). Most cracks appear filled with a grey-blueish phase, which together with the

Cu enrichment observed in the EDS maps, is interpreted as Cu-oxide (CuOx). This indicates

Manuscript III

111

that the local pO2 has been sufficiently high to oxidize metallic Cu in these areas. All cracks

extending to the surface are in direct connection with a Cu/CuOx “particle” in the OOZ.

Most remarkable in this respect is the crack branching into three cracks close to the sample

corner in Fig. 7.3-b (right). All three CuOx-filled branches end in cauliflower-like metallic

Cu at the surface.

Fig. 7.3. (a) Secondary electron micrograph of the top surface morphology of the oxidized

(Zr55Cu30Al10Ni5)98Er2 BMG; (b) cross-sectional light optical microscopy; (c) Cross-sectional back-scatter

electron micrograph of the oxide zones developed in the surface of the BMG and corresponding EDS

elemental mapping.


112

In some areas in IOZ, a branching of cracks can also be observed. Such branching could be

a consequence of the volume expansion associated with the conversion of Cu into CuOx

inside the cracks. The development of micro-cracks filled with Cu has been reported and

discussed in our previous works [23–24]. The discontinuities in the transition from zone I

to II at the location of a crack are evidence that the cracks formed after the formation of the

IOZ.

Transmission electron microscopy

TEM investigations were carried out to obtain additional information on the various

crystalline phases that have developed in the oxidation zones of the BMG after the in-situ

XRD experiment. TEM lamella preparation was conducted on top of the area where a cap

of the mushroom-like oxide feature was observed in the surface region. Fig. 7.4-a presents

an overview of a ~ 5 μm lamella extracted along the surface normal, showing the oxidized

regions. TEM bright-field (BF) imaging combined with the average EDS results obtained

from the areas in Fig. 7.4 a-b marked as “A”, “B” and “C” are provided. The results show

that the OOZ (A) is enriched in copper and oxygen, while the IOZ (B) mainly consists of

Zr and O. The chemical compositions obtained from “A” and “B” are consistent with Cu-

based oxides and a mixed oxide zone, respectively. The mixed oxide zones in the IOZ

contains t-ZrO2 and Cu2O where also the other BMG alloying constituents are detected,

albeit strongly depleted in Cu. This strongly supports the results presented in Figs. 7.1-3

where the OOZ consists of (predominantly) CuO and the IOZ contains mostly t-ZrO2.

Moreover, the magnified BF-TEM image of an “area 1” (designated with dotted circle in

Fig. 7.4-a) is given in Fig. 7.4-b, demonstrating a Cu-enriched micro-crack directed to the

surface. The Cu enrichment in zone “C” is confirmed by the average point-to-point EDS

analysis together with the elemental mapping obtained from the dotted pink squared area

and presented in the right-hand side of Fig. 7.4. This decorated line is described as a Cu-

enriched self-healed micro-crack. The occurrence of this surprising phenomenon was

observed and reported for other BMGs in Refs.[23–24, 26] . In addition to TEM BF cross-

sectional imaging, to analyze the crystal structure of the phases present in the oxidized

zones and comparing the results with XRD, selected area electron diffraction (SAED) was

performed on positions “1” and “2” labeled in Fig. 7.4-a. SAED patterns provided in Fig.

Manuscript III

113

7.4 c-d confirm the presence of t-ZrO2 and Cu2O in IOZ and monoclinic CuO in OOZ,

respectively. This is consistent with the obtained results from the XRD for phase analysis

(Fig. 7.1-b and Fig. 7.2). The SAED pattern obtained from “position 2” indicates that it is

acquired from a single CuO crystal, while ring patterns obtained from “position 1” indicates

diffraction from many grains and, hence, nano-crystallinity of ZrO2 and Cu2O. Moreover,

based on the EDS chemical composition results, Al is also present in the mixed oxide zone

(IOZ) together with t-ZrO2, but no diffraction spots of Al-oxide were found (Fig. 7.4-c).

This is in agreement with the ex-situ and in-situ XRD results. It is anticipated that Al in the

IOZ is dissolved in t-ZrO2. This anticipation has also been previously stated in Refs. [27–

28], where no indication of Al2O3 in IOZ was observed.

Fig. 7.4. (a) TEM bright field micrograph of the (Zr55Cu30Al10Ni5)98Er2 BMG air-oxidized during in-situ XRD

experiment at 600 K for 60 hr together with the average point EDS analysis from the specified zones (A, B,

and C). (b) Enlarged image of the selected area 1 in (a). SAED patterns of position (c) 1 and (d) 2. The EDS

elemental mapping from the pink selected area in Fig. 4-b is presented on the right.

Evaluation of the lattice strain and residual stress profiles

In-situ conventional lattice strain XRD sin2ψ investigation was performed to non-

destructively monitor the evolution and (partial) relaxation of stresses in various crystalline


114

phases developing in the surface-adjacent region of the BMG during air-oxidation4. Based

on this method, the lattice spacing is affected by the growth stresses associated with

oxidation of the BMG. The elastic lattice strain can be converted to residual stresses

applying the material’s properties [29].

It was previously demonstrated that the state of stress in the surface of the BMG during

oxidation is biaxial rotationally symmetric (σ11=σ22=σ//) [26]. Hence, the elastic lattice

strain ( ,

hkl

) obeys [29]:

, 0 2

, 2 / / 1 / /

0

1sin 2

2

hkl hkl

hkl hkl hkl

hkl

d dS S

d

7-1

where ,

hkl

is the elastic strain in a particular direction defined by the rotation and tilt

angles, φ and ψ, respectively. ,

hkld and 0

hkld are the strained and strain-free interplanar

lattice spacing, respectively. 1

hklS and 21/ 2 hklS are the X-ray elastic constants (XECs). In

order to calculate the XECs of tetragonal ZrO2, monoclinic CuO, and cubic Cu2O, the Voigt

approximation was applied [30]. Table 7.1 provides the elastic constants of t-ZrO2, m-CuO

and c-Cu2O including the single crystal elastic coefficients (C), XECs, Poisson ratio (ν) and

Young’s modulus (E).

The interplanar lattice spacing (011

,d ) was then calculated and plotted as a function of sin2ψ.

The strain-free direction follows from Eq. 7-1 and is given by 2 10

2

2sin

(1/ 2)

hkl

hkl

S

S . The

average stress values (σ//) were deduced from the slope of the linear regression through the

4 It is noted that the intensity of the radiation used in a conventional laboratory XRD is not

brilliant enough to determine the state of stress over the whole thickness of the oxidation

region, specially where it might be compositional and stress gradient through the oxidation

zones. However, it is sufficient to provide a picture regarding the development of the

residual stresses in the multiple crystalline oxidation zones in the case of this study.

Manuscript III

115

data points, applying the appropriate XECs. The average macro-stress (σ//) follows from

,

/ / 2

2 0

2

sin

hkl

hkl hkl

d

S d

(cf. Eq. 1). The standard deviation of a linear fit was taken as an

error estimate for the obtained stress values.

Table. 7.1. The single crystal elastic coefficients [28–29], calculated X-ray elastic constants (XECs)

based on Voigt assumption, Poisson’s ratio (ν) and Young’s modulus (E) of t-ZrO2 [33], m-CuO5

and c-Cu2O [34] are summarized, respectively.

C11

(GPa) C12 C13 C33 C44 C66

S1

(10-6 MPa-1)

1/2S2

(10-6 MPa-1) ν

E

(GPa)

t-ZrO2 327 100 62 264 59 64 -1.28 6.05 0.32 192

m-CuO 196.41 122.63 114.64 293.7 20.28 56.44 -3.69 0.132 0.39 89.6

c-Cu2O 127.46 107.07 - - 9.21 - -0.164 0.521 0.45 30

The evolution of residual stresses in the surface region of the BMG exposed to the XRD

sin2ψ air-oxidation during heating, isothermal holding and cooling stages for the reflections

of (011) t-ZrO2, (111) m-CuO and (111) c-Cu2O as a function of time are presented in Fig.

7.5. The error bars reflect twice the standard deviation obtained on linear regression of the

dhkl-sin2ψ data. Based on the negative sign of the stress, all oxide phases are in compression.

On heating, at temperature ~ 550 K6, the conversion of Zr to t-ZrO2 evokes in-plane

compressive residual stresses of -1430±120 MPa within the surface of the BMG. On

5 In this study, in order to calculate XECs of monoclinic CuO (α=γ=90˚ and β≠90˚), it has

been assumed that it has trigonal lattice structure (a=b, α=β=90º and γ=120º) which is a

limiting case of a simple monoclinic Bravais lattice with β=120˚ [35].

6 A set of diffractograms acquired at temperatures ≥ 550 K was used for the calculation of

lattice strains, even though the onset temperature of t-ZrO2 formation in the surface region

of the investigated BMG detected during in-situ XRD air-oxidation experiment is ~ 470 K

using Cr-Kα radiation. It is due to the lack of intensity of (011) t-ZrO2 reflection compared

with the amorphous hump probed by the applied Cu-Kα radiation, where a reliable

diffraction peak fitting could not be applied for the diffractograms obtained at temperatures

≤ 550 K. It should also be noted that the temperature where the first t-ZrO2 was detected


116

continued heating to 600 K, compressive stresses are largely relaxed and the average stress

value decreases to -630±500 MPa and later to -270±120 MPa. The large standard deviation

of the stress values observed during heating with further exposure time could indicate the

presence of different stress states in the analyzed volume at the onset of stress relaxation.

During isothermal holding at 600 K for 60 hr, the developing compressive residual stresses

in t-ZrO2 remain almost stable at an average value of -440±90 MPa. On cooling to ambient

temperature a slight increase in mean compressive stress as compared with the isothermal

stage, -550±60 MPa is observed, which implies that the underlying BMG substrate shrinks

more than the surface oxidized region. This indicates the contribution of the thermal

stresses as a consequence of the thermal expansion mismatch between the BMG and the

oxidation zones, which is consequently, superimposed to the growth stresses during

cooling.

The average growth compressive stresses caused by the volume expansion as a result of

Cu-based oxides formation in the OOZ during isothermal heating and measured for the

reflections (111) m-CuO and (111) c-Cu2O are -100±20 MPa and -50±10 MPa,

respectively. Obviously, compressive stresses in the Cu-oxides are not as significant as the

stresses developed in t-ZrO2. Moreover, a tendency towards larger compressive stress

values during cooling is revealed for m-CuO and c-Cu2O phases, which could also be

attributed to additional thermal stress. It is concluded that thermal shrink of the un-oxidized

BMG substrate is more pronounced than the thermal shrink of the crystalline oxides formed

in the IOZ and OOZ. Evidently, the largest discrepancy from the thermal expansion

coefficient of the BMG is observed for c-Cu2O. This would be consistent with a negative

thermal expansion coefficient for this phase (cf. lattice expansion for c-Cu2O on cooling in

Fig. 7.2).

by the applied Cu Kα radiation with larger information depth compared to Cr-Kα radiation

was ~ 510K.

Manuscript III

117

Fig. 7.5. Residual stress vs. oxidation time obtained from in-situ lattice strain XRD sin2ψ method during air-

oxidation of (Zr55Cu30Al10Ni5)98Er2 BMG, consisting of heating, isothermal and cooling stages for {011} t-

ZrO2,{111} m-CuO and {111} c-Cu2O phases. At the top, the plot also shows the thermal cycle applied

during in-situ lattice strain XRD sin2ψ measurements.

Discussion

118

7.4 Discussion

In the current study, oxidation of (Zr55Cu30Al10Ni5)98Er2 BMG in air at temperatures lower

than 600 K was investigated in-situ in order to reveal insight in the microstructure evolving

during oxidation and in particular the role of internal (residual) stress in the crystalline

oxides. The evolution of the microstructure, stress relaxation mechanisms and stress-

assisted self-healing mechanism during long-term sub-Tg exposure of

(Zr55Cu30Al10Ni5)98Er2 BMG to an oxygen-containing atmosphere is summarized in Fig.

7.6.

Evidently, the reaction of the material with oxygen present in the atmosphere starts with

the formation of small crystals of t-ZrO2 inside the BMG. Inevitably, bonding of Zr with

oxygen is accompanied by a volume expansion of approximately 40%. Since the ZrO2

forms as nano-crystals inside the BMG under the ingress of O into the BMG, these stresses

cannot be relaxed by additional growth in a direction perpendicular to the surface. Thus,

large compressive stresses build up in the IOZ, which are compensated by tensile stresses

in the underlying (and surrounding) BMG. Initially, these compressive stresses (in ZrO2)

are so large (1.4 GPa in average, see Fig. 7.5) that (locally) the elastic limit of the material

is exceeded; for the BMG the yield stress is about 1.5 GPa. The reaction of the material to

this volume expansion and associated stress build-up is the rejection of Cu from the

internally oxidized part of the BMG. Evidently, Cu segregates to free surfaces, either the

original surface of the BMG or surfaces created by cracking of the BMG. The driving force

for the redistribution and outward migration of Cu (and Ni), which does (do) not dissolve

in t-ZrO2, is the compressive stress, i.e. stress-induced diffusion.

Rejection of metallic Cu might explain the observed stress relaxation (from ~1.4 GPa to

~0.3 GPa) during heating of the material to 600 K, the temperature of isothermal oxidation

treatment. Nevertheless, a lower coefficient of thermal expansion (CTE) of t-ZrO2 (surface)

as compared to the CTE of the BMG (bulk) also contributes to relaxation of the (average)

state of compressive stress in the IOZ. Most importantly, the formation of cracks in the IOZ

leads to stress relaxation. Cracks formed by compressive stresses are expected to be parallel

to the plane of the surface. In the present investigations the occurrence of cracks

perpendicular to the surface (Fig. 7.3) indicates that, at least locally, tensile stresses were

present during crack initiation. It is argued that crack initiation occurred in the BMG

Manuscript III

119

adjacent to the IOZ as a consequence of a compensating tensile stress in this part of the

BMG, in an early stage of oxidation.

Upon the initiation of cracks new locations for Cu segregation are created. Oxidation of

metallic Cu that has segregated to the surface or crack surfaces can only happen if the

partial pressure of oxygen is higher than 10-22 bar, i.e. the partial pressure necessary to

oxidize Cu to Cu2O [26] . Oxidation of Cu is expected to cause compressive stresses in

oxides in the OOZ. These stresses are modest, close to negligible, because at the outer

surface, relaxation of growth stresses can easily be achieved by additional growth of the

oxide in the direction perpendicular to the surface. The growth stresses associated with the

formation of Cu-oxides inside the cracks leads to wedge formation inside the crack and

may thus lead to crack growth. It is believed that the cracks extending deep into the IOZ

are a consequence of early crack-initiation in the BMG and crack-growth on continued

oxidation by wedge formation inside the crack. The self-repaired cracks perpendicular to

the surface in the outer part of the IOZ show that many of the initiated cracks are pacified

by the self-repair. The observation of a discontinuity in the zone I to zone II transition at

the location of a crack (Fig. 7.3-c) suggests that crack growth by CuO-wedge formation

occurs at a later stage of oxidation.

On continued growth of the IOZ and the OOZ at 600 K under the inward diffusion of O

and outward diffusion of Cu (and Ni), respectively, the outer part of the IOZ becomes

depleted in Cu. The associated reduction of compressive stress implies that this part of the

IOZ7 may be the impetus for wedge formation.

The above discussion demonstrates that more detailed investigations may be necessary to

reveal all mechanisms involved in stress development, stress relaxation and self-healing

during oxidation of BMGs.

7 Note that the stress determination with XRD relies on lattice strain in ZrO2 only. The

overall state of stress in the entire IOZ can be different.

Conclusion

120

Fig. 7.6. An overview of the proposed oxidation mechanism and the compressive stress-assisted self-healing

phenomenon of oxidized (Zr55Cu30Al10Ni5)98Er2 BMG.

7.5 Conclusion

The present study deals with in-situ X-ray diffraction, monitoring the lattice strains

(residual stresses) developing in crystalline oxides that form during oxidation of

(Zr55Cu30Al10Ni5)98Er2 BMG below Tg at a 600K. Initially, under the ingress of oxygen the

BMG develops nano-crystalline tetragonal ZrO2 at a temperature as low as ~470 K,

reflecting the very high affinity of Zr for oxygen. The volume expansion associated with

ZrO2 formation inside the internal oxidation zone (IOZ) is constrained by the BMG

substrate and consequently a noticeable compressive residual stresses of ~ -1.4 GPa

develops. The high compressive stress induces a driving force for the diffusion of Cu-atoms

(and most likely also Ni), which subsequently segregates at the surface. The appearance of

Cu at the surface coincides with a reduction of the compressive stresses in t-ZrO2 from ~ -

1.4 GPa to ~ -400 MPa during oxidation. Although Cu-redistribution is anticipated to be

accompanied by a reduction of the stress in the IOZ, perhaps more importantly, crack-

initiation in the BMG adjacent to the IOZ as a consequence of compensating tensile stresses

is held responsible for the observed stress relaxation. These cracks form perpendicular to

the surface. The creation of crack surfaces leads to new locations for the segregation of Cu

and consequently self-healing of the crack. Subsequent oxidation of Cu at the surface forms

the outer oxidation zone (OOZ). This surface oxide hardly experiences residual stresses, as

they are relaxed during growth of the Cu-oxide perpendicular to the surface. Cu-oxide

Manuscript III

121

development inside the cracks forms a wedge and can lead to crack growth into the IOZ.

Cu-redistribution leads to an important reduction of the Cu-contents in the IOZ.

Acknowledgements

The financial support from Villum Fonden (Grant No. 13253) is sincerely acknowledged

during this research was conducted. The authors would also like to thank the Institute of

Applied Physics, Jiangxi Academy of Sciences, Nanchang, 330029, China for providing

the as-cast plate of (Zr55Cu30Al10Ni5)98Er2 BMG.

Conclusion

122

References

[1] W. L. Johnson, “Thermodynamic and kinetic aspects of the crystal to glass

transformation in metallic materials,” Prog. Mater. Sci., vol. 30, no. 2, pp. 81–134,

1986.

[2] G. Tarjus, “Metal turned to glass,” Nature, vol. 448, no. 7155, pp. 758–759, 2007.

[3] C. J. Byrne and M. Eldrup, “Bulk metallic glasses,” Science (80-. )., vol. 321, no.

5888, pp. 502–503, 2008.

[4] J. F. Löffler, “Bulk metallic glasses,” Intermetallics, vol. 11, no. 6, pp. 529–540,

2003.


vol. 54, no. 3, pp. 321–326, 2006.

[6] M. Chen, “A brief overview of bulk metallic glasses,” NPG Asia Mater., vol. 3, no.

9, pp. 82–90, 2011.

[7] W. L. Johnson, “Bulk amorphous metal-An emerging engineering material,” Jom,

vol. 54, no. 3, pp. 40–43, 2002.

[8] A. I. Salimon, M. F. Ashby, Y. Bréchet, and A. L. Greer, “Bulk metallic glasses:

What are they good for?,” Mater. Sci. Eng. A, vol. 375–377, no. 1–2 SPEC. ISS.,

pp. 385–388, 2004.

[9] W. H. Wang, “Bulk metallic glasses with functional physical properties,” Adv.

Mater., vol. 21, no. 45, pp. 4524–4544, 2009.

[10] L. Huang et al., “A Zr-based bulk metallic glass for future stent applications:

Materials properties, finite element modeling, and in vitro human vascular cell

response,” Acta Biomater., vol. 25, pp. 356–368, 2015.

[11] A. Takeuchi and A. Inoue, “Classification of Bulk Metallic Glasses by Atomic Size


to Characterization of the Main Alloying Element,” Mater. Trans., vol. 46, no. 12,

pp. 2817–2829, 2005.

[12] M. K. Miller and P. Liaw, Bulk Metallic Glasses: An Overview. 2007.

[13] M. Chen, “Mechanical Behavior of Metallic Glasses: Microscopic Understanding of

Strength and Ductility,” Annu. Rev. Mater. Res., vol. 38, no. 1, pp. 445–469, 2008.


[15] K. Mondal, U. K. Chatterjee, and B. S. Murty, “Oxidation behavior of

multicomponent Zr-based amorphous alloys,” J. Alloys Compd., vol. 433, no. 1–2,

pp. 162–170, 2007.



[17] Z. Xu, Y. Xu, A. Zhang, J. Wang, and Z. Wang, “Oxidation of amorphous alloys,”

J. Mater. Sci. Technol., vol. 34, no. 11, pp. 1977–2005, 2018.

Manuscript III

123

[18] S. Stanojevic, I. Gallino, H. Aboulfadl, M. Sahin, F. Mücklich, and R. Busch,

“Oxidation of glassy Ni-Nb-Sn alloys and its influence on the thermodynamics and

kinetics of crystallization,” Acta Mater., vol. 102, pp. 176–186, 2016.

[19] C. Y. Tam and C. H. Shek, “Oxidation-induced copper segregation in Cu60Zr30Ti10

bulk metallic glass,” J. Mater. Res., vol. 21, no. 4, pp. 851–855, 2006.

[20] K. Weller, L. P. H. Jeurgens, Z. Wang, and E. J. Mittemeijer, “Thermal oxidation of

amorphous Al0.44Zr0.56 alloys,” Acta Mater., vol. 87, pp. 187–200, 2015.

[21] W. Kai et al., “Air-oxidation of a Zr50Cu43Al7 bulk metallic glass at 400-500°C,”

Corros. Sci., vol. 64, pp. 98–104, 2012.

[22] S. Lu, S. Sun, G. Tu, X. Huang, and K. Li, “The effect of yttrium addition on the air

oxidation behavior of Zr-Cu-Ni-Al bulk metallic glasses at 400–500 °C,” Corros.

Sci., vol. 137, no. March, pp. 53–61, 2018.

[23] S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, and M. A. J. Somers,

“Surface hardening by gaseous oxidizing of (Zr55Cu30Al10Ni5)98Er2 bulk-metallic

glass,” J. Alloys Compd., vol. 800, pp. 456–461, 2019.




[25] U. Köster, L. Jastrow, and M. Meuris, “Oxidation of Cu60Zr30Ti10 metallic glasses,”

Mater. Sci. Eng. A, vol. 448–451, pp. 165–168, 2007.

[26] S. Haratian et al., “Strain, stress and stress relaxation in oxidized ZrCuAl-based bulk

metallic glass.”

[27] S. Inamura, H. Miyamoto, Y. Imaida, M. Takagawa, K. Hirota, and O. Yamaguchi,

“Formation and hot isostatic pressing of ZrO2 solid solution in the system ZrO2-

Al2O3,” J. Mater. Sci., vol. 29, no. 18, pp. 4913–4917, 1994.

[28] S. Moreau, M. Gervais, and A. Douy, “Formation of metastable solid solution in the

ZrO2 rich part of the system ZrO2-Al2O3.” pp. 625–631, 1997.


Evaluation-Application-Assessment. 1997.

[30] P. Schoderbock, P. Leibenguth, and M. Tkadletz, “Pattern decomposition for

residual stress analysis: A generalization taking into consideration elastic anisotropy

and extension to higher-symmetry Laue classes,” J. Appl. Crystallogr., vol. 50, pp.

1011–1020, 2017.

[31] E. H. Kisi and C. J. Howard, “Elastic Constants of Tetragonal Zirconia Measured by

a New Powder Diffraction Technique,” J. Am. Ceram. Soc., vol. 81, no. 6, pp. 1682–

1684, 1998.

[32] N. Liu, J. Sun, and D. Wu, “Elastic constants and thermodynamic properties of Cu,

Cu2O and CuO from first-principles calculation,” Adv. Mater. Res., vol. 335–336,

pp. 328–332, 2011.

[33] X. S. Zhao, S. L. Shang, Z. K. Liu, and J. Y. Shen, “Elastic properties of cubic,

Conclusion

124

tetragonal and monoclinic ZrO2 from first-principles calculations,” J. Nucl. Mater.,

vol. 415, no. 1, pp. 13–17, 2011.

[34] A. Altaweel, T. Gries, S. Migot, P. Boulet, A. Mézin, and T. Belmonte, “Localised

growth of CuO nanowires by micro-afterglow oxidation at atmospheric pressure:

Investigation of the role of stress,” Surf. Coatings Technol., vol. 305, pp. 254–263,

2016.

[35] M. J. Mehl et al., “The AFLOW Library of Crystallographic Prototypes: Part 1,”

Comput. Mater. Sci., vol. 136, pp. S1–S828, 2017.

Manuscript IV

125

8 Manuscript IV

Strain, stress and stress relaxation in oxidized ZrCuAl-based bulk

metallic glass1

Saber Haratiana,b, Frank Niessenb, Flemming B. Grumsena, Mitchell J. B. Nancarrowb,

Elena Perelomab, Matteo Villaa, Thomas L. Christiansena, Marcel A.J. Somersa

a Materials and Surface Engineering Section, Mechanical Engineering Department,


Denmark

b Electron Microscopy Centre, University of Wollongong, Wollongong, New South Wales

2500, Australia

Abstract

Surface engineering of Zr51.3Al8.5Cu31.3Ni4Ti4.9 bulk metallic glass (BMG) by gaseous

oxidizing below the glass-transition temperature is investigated as a means to introduce

compressive residual stress in the surface region. The ZrCuAl-based BMG was exposed to

an extremely low oxygen partial pressure of 10-41 bar at 600 K for 60 h. The oxidizing

treatment led to the formation of an internal oxidation zone, consisting of finely dispersed

crystalline cubic ZrO2 (c-ZrO2), metallic regions inclined with the surface and Cu-hillocks

at the surface. The stresses introduced by the volume expansion associated with oxidation

were evaluated from i) the lattice strains within c-ZrO2, as determined with an X-ray

diffraction (XRD) based method, and ii) strain-relaxation as a response to annular focused

ion beam (FIB) milling, as monitored with digital image correlation (DIC). XRD analysis

yielded -1.5 GPa (compressive stress) in the crystalline c-ZrO2, while the strain relaxation

monitored with FIB-DIC analysis indicated compressive residual stresses of -1.4 GPa in

the internal oxidation zone. The strain and stresses determined with the independent

measurement methods are discussed. The quantitative macro-strains are discussed in

relation to the microstructural features and stress relaxation mechanisms during evolution

of the internal oxidation zone.

1 Unpublished work at the date of Ph.D. thesis submission. The final article may deviate

from the present manuscript.

Introduction

126

8.1 Introduction

The discovery of metallic glasses (MGs), i.e. metals without a periodic crystalline structure,

has obtained considerable attention since 1960 [1–2]. MGs have been developed and

synthesized in several multi-component alloying systems in a bulk form with a minimum

diameter (or thickness) of 1 mm, which represents a new class of advanced materials [3].

The lack of an ordered atomic arrangement is reflected in the absence of strength-limiting

structural defects, which in common metals are defined by interfaces and dislocations [4–

5]. Essentially, BMGs are attractive engineering materials with competitive (and often

unique) mechanical and physical properties [6–7]. Their outstanding properties, i.e. highest

yield strength and elastic limit among all metals, has aroused technological interest [8–11].

Unfortunately, the attractive outstanding elastic properties are outweighed by shortcomings

in plastic deformation, where strain softening limits plastic deformation, especially when

subjected to tensile loading. Due to the absence of interfaces within the amorphous

structure, the plastic flow in BMGs is only concentrated in localized shear bands and their

unhindered propagation ultimately results in catastrophic failure, as soon as their critical

length is reached [12–15]. Therefore, attention has turned to the development of strategies

to effectively mitigate the propagation of shear bands and thereby extend the plasticity of

BMGs. Several systematic studies provided effective processing techniques aimed at

improving the plasticity and fracture toughness [16–20]. Among other approaches,

introduction of a state of compressive residual stresses in the surface region of BMGs by

surface engineering was proposed to prevent crack initiation [21–23]. Such compressive

stresses could counteract local stress concentration to prevent crack initiation and thus

improve the performance of a BMG during service exposure.

Recent investigations suggest that surface hardening by low-temperature gaseous

oxidizing of ZrCuAl-based BMG in a controlled atmosphere is feasible [24]. Applying a

high oxygen partial pressure (pO2), it was demonstrated that the dissolution of oxygen into

the surface zone of the BMG results in the formation of internal and outer oxidation zones.

In the internal oxidation zone (IOZ) the surface hardness reaches up to ~1200 HV by the

formation of nano-crystalline zirconia (ZrO2). The outer oxidation zone (OOZ) consists of

Cu-based oxides. Moreover, a surprising self-repair phenomenon was observed. The

mechanism of self-repair was suggested to be stress-induced diffusion of noble elements to

micro-crack surfaces (as well as the BMG surface, where it forms the OOZ), where they

Manuscript IV

127

segregate as crystalline metals and effectively seal the crack [25]. So far, the actual

development of compressive residual stresses and associated macro-strains in the hardened

case of the oxidized BMG has not been demonstrated. For this purpose, the present

experimental study seeks to experimentally determine the macro-strains and macro-stresses

in thermochemically surface-engineered BMG, to arrive at better understanding the

microstructural evolution and the mechanism behind the self-healing phenomenon,

observed on oxidation of multi-component noble metal containing Zr-based BMGs. Here,

oxidizing in an extremely low pO2 is performed to suppress the development of the OOZ.

In order to assess the macro-strains, two methods were utilized. For the determination of

lattice strains in crystalline c-ZrO2 angular dispersive X-ray diffraction (XRD) was utilized.

The XRD sin2ψ method is among the most widely applied and most reliable techniques for

evaluating (elastic) residual stresses in crystalline phases from lattice strains [26–27]. The

elastic strain in a crystalline material changes the interplanar spacing for a specific

reflection (hkl) ( hkld ) from its unstressed value ( 0

hkld ). Generally, for thermochemically

surface-engineered materials a state of bi-axial stress can be assumed. The change in lattice

spacing, , 0

hkl hkld d , depends on the rotation angle, , and the tilt angle, , of the diffraction

vector with respect to the system of principal stresses (cf. Fig. 8.3 in the Results section).

The determined lattice strain is an average over all diffracting grains and can be converted

into an average elastic residual stress after adopting a grain-interaction model [26] (see

further in the Results section).

Additionally, strain relaxation as monitored with digital image correlation (DIC) in

response to incremental focused ion-beam (FIB) milling [28–29] was used as a

complementary method for evaluating the local strain relief in the entire IOZ (not only in

crystalline c-ZrO2). DIC is applied for monitoring the displacements caused by strain relief

following incremental ring-core FIB milling in the surface region [30] and was firstly

proposed by Korsunsky et al. [31]. Essentially, the method is a miniaturization of the

macroscopic ring-core method developed by Keil in 1992 [32]. In this technique, local

material removal within a well-defined gauge volume, using high-energy ions, induces

relaxation of the strain fields associated with residual stresses while the resulting

displacements are tracked by DIC [33]. This method enables the determination of the

average strain in the internal oxidation zone, including crystalline and amorphous phases.

Experimental details and procedures

128

FIB-DIC based techniques can be used to evaluate stresses in areas of a few squared

micrometers [34], while XRD provides information over a much larger surface area.

8.2 Experimental details and procedures

Sample preparation and characterization

The investigations were performed on as-cast Zr51.3Cu31.3Al8.5Ni4Ti4.9 (at.%) BMG

synthesized using vacuum arc-melting of high purity (99.9 wt. %) elemental constituents

in a Ti-gettered argon atmosphere. The details of the preparation procedure can be found

elsewhere [24–25]. Differential thermal analysis in a Netzsch STA 449C thermal analyzer

under a 50 cm3.min-1 flow of argon was used to determine the glass transition temperature

(Tg= 670 K) and the onset crystallization temperature (Tx= 705 K). BMG samples of 3×3×1

mm3 were thermochemically exposed to oxidizing at 600 K for 60 h in the thermal analyzer

(Netzsch STA 449C). The applied gas mixture imposes an extremely low oxygen partial

pressure (pO2≈ 10-41 bar) which was realized chemically at a total gas pressure of 1 bar. An

extremely low pO2 was applied to avoid the formation of a porous non-homogenous outer

oxidized network as was observed in previous studies at higher pO2 [24–25]. Prior to

oxidizing, the sample surfaces were carefully polished to 1 µm diamond and cleaned with

ethanol.

Phase analysis of the amorphous and thermochemically oxidized BMG samples was

conducted with X-ray diffractometry, using a Bruker D8 Discover equipped with CuKα

radiation (λ=0.15418 nm). For all measurements, a scattering angle, 2θ, range of 25°-90°,

a step size of 0.04° and an acquisition time of 8 s per step were applied. The surface

topography of the oxidized BMG was characterized using a FEI Helios NanoLAB G3 CX

dual-beam scanning electron microscope (SEM) equipped with an EDAX SD Apollo 10

Pegasus energy dispersive X-ray spectroscopy (EDS) system. Furthermore, the cross-

section of the treated sample was examined with a Neophot 32 (Zeiss, Jena) reflected light

optical microscope (LOM) and a Supra 35 SEM (Carl Zeiss) fitted with a Thermo NORAN

system 6 EDS detector.

XRD lattice strain determination

In this study, ex-situ XRD lattice strain analysis was carried out with a Bruker D8

diffractometer equipped with Cu-radiation generated at 40 kV and 40 mA. The lattice

spacing for residual stress analysis of the thermochemically oxidized BMG was determined

Manuscript IV

129

in lock-coupled parallel beam incident geometry using a point focus configuration

(collimator size: 1 mm). The measurements were performed in the 2θ range 27°-33°. This

range was selected to include the 111 reflection of c-ZrO22. A step size (∆2θ) of 0.03° and

a counting time per step 10 s were chosen for acquiring diffractograms at seven side-

inclination (ψ) angles (0°, 10°, 20°, 30°, 40°, 50°, and 60°) and four rotation () angles (

0°, 90°, 180°, and 270°). The peak fitting of the X-ray diffractograms was performed using

a Pseudo-Voigt function with the TOPAS P2-1 software (Bruker AXS).

FIB-DIC measurement

In the experimental approach, the ring-core milling geometry was found to be advantageous

as it provides a progressive homogenous strain relief in all directions at the surface region

[35]. The strain relaxation was digitally monitored in post by tracing deposited

high-contrast features in high-resolution SEM images acquired during the milling process.

In this work, FIB milling was conducted using a FEI Helios NanoLAB G3 CX dual-beam

FIB/SEM microscope. The method was automated using the AutoScript 4 plugin and the

FIB milling parameters were optimized for the specific material. A speckle pattern was

deposited onto the sample surface by electron-beam assisted platinum (Pt) deposition at 10

kV and 1.4 nA. The pattern was applied using exact numerically controlled electron-beam

positioning via stream files [36] and comprises a uniform, non-periodic array of Pt dots

[37]. The contrast-rich features are required for the DIC algorithm to track the local

displacements associated with surface strain relaxation. The amount and size of the

speckles was scaled with respect to the pillar diameter. The regions of the oxidized surface

selected for investigation with the milling process were free of copper particles (cf. Fig.

8.2), as to minimize Cu-redeposition. For incremental ring-core milling, an annular trench

was milled around the Pt-deposited pattern, using a focused Ga+ beam with 30 kV

accelerating voltage and 230 pA current. The outer diameter of a ring-core geometry was

set to twice the milling depth plus the cylinder diameter (2h+D) to provide sufficient space

for the sputtered material. Each single-pass milling step was carried out at a fixed dwell

2 Other reflections of c-ZrO2 were not selected for the lattice strain measurement, because

of low diffracted intensity and/or overlap with metallic copper peaks.

Experimental details and procedures

130

time of 150 µs per pixel. The milling direction was set from the outer to the inner side of

the ring-core to minimize re-deposition of sputtered material onto the cylinder. Before and

after each milling cycle, a series of high-resolution SEM images of the speckle pattern was

acquired. The brightness and contrast of the SEM images were adjusted to evenly distribute

the pixel intensities from 0% to 90% of the available histogram to keep a safety margin in

case of contrast changes during the acquisition. The adjusted imaging parameters were kept

constant throughout the milling. The auto-drift image correction function in the FIB/SEM

microscope was used to minimize the direct effect of a slight electron beam drift during

SEM imaging on the strain relief results. The FIB-DIC process parameters used in this work

are summarized in Table 8.1.

Table 8.1. The FIB-DIC process parameters optimized for the application of the method to c-ZrO2 formed in

the surface adjacent region of Zr51.3Cu31.3Al8.5Ni4Ti4.9 BMG.

Optimized process parameters

Electron beam

assisted surface

speckle

decoration

Application Pt deposition

Accelerating voltage (kV) 10

Beam current (nA) 1.4

Number of speckles (4, 6, and 8 μm pillar) 160, 200, and 205

High-resolution

SEM imaging



Detector Through the lens detector

(TLD)

Image resolution (pixels) 1536×1024

Image acquisition Integration of 128 images at a

dwell time of 50 ns*

Sample tilt angle (°) 52

Tilt correction Activated**

Working distance (mm) 3.9

Focused ion beam

milling

Application (non-gas assisted milling) Silicon



Milling direction Outer to inner

Number of passes 1

Dwell time (μs) 150

Pillar size (inner-outer diameter (μm) / Number

of milling cycles)

4-14 / 80

6-18 / 100

8-24 / 185

* Image integration was applied to reduce the scanning noise in the image, which may generate ghost-strains

in the DIC analysis.

** Tilt correction was applied to achieve an identical scanning resolution in the X and Y direction of the

image.

The DIC algorithm is based on the function cpcorr.m of the MATLAB Image Processing

Toolbox [38]. To correct for potential drift of the SEM images low-resolution DIC was

tracked using a coarse grid of sub-images (facets).

Manuscript IV

131


X-ray diffraction and microscopical characterization

X-ray diffractograms of the as-cast and the thermochemically oxidized

Zr51.3Cu31.3Al8.5Ni4Ti4.9 (at. %) BMG samples are shown in Fig. 8.1. For the as-cast

material, the two broad intensity maxima at 2θ ≈ 38º and 65º show that no crystalline

compounds are present within the information depth probed by the applied X-rays, but that

the sample is amorphous. Oxidizing for 60 h at 600 K in an extremely low pO2 leads to the

formation of c-ZrO2 in the surface adjacent region of the BMG. In addition to c-ZrO2,

diffraction peaks of metallic copper, which (partly) overlap with c-ZrO2 peaks, may be

present. No copper oxide phases developed.

Fig. 8.1. The X-ray diffractograms of the as-cast and thermochemically oxidized Zr51.3Al8.5Cu31.3Ni4Ti4.9

BMGs.

The results of microscopical investigations of the thermochemically oxidized

Zr51.3Al8.5Cu31.3Ni4Ti4.9 BMG at 600 K for 60 h in a low pO2 are given in Fig. 8.2. The

secondary electron (SE) micrograph of the surface morphology of the oxidized BMG (Fig.

8.2-a) reveals a grey surface containing finely distributed, facetted particles (one is marked

by a white arrow). The chemical compositions of the grey surface and the particles were

investigated using energy dispersive X-ray spectroscopy (EDS). Point-to-point EDS results

(not provided) showed that the grey surface contains all BMG alloying elements (Zr, Cu,

Al, Ni, and Ti) and oxygen. The facetted individual particles are pure metallic copper; all

other elements are absent. This would be consistent with the presence of Cu peaks in the


132

diffractogram (Fig. 8.1). The cross-section in Fig. 8.2-b shows the presence of a 9-µm-thick

internal oxidation zone (IOZ), which appears dark in light optical microscopy and

terminates abruptly. The light-contrast lines are inclined about 60° with the plane of the

surface and indicate that these decorated regions contain a metallic phase, most likely Cu

(see [25]), which, provided that these lines are crystalline, would again be consistent with

the presence of Cu-peaks in Fig. 8.1. The presence of a metallic phase in the IOZ is further

confirmed by the back-scatter electron (BSE) contrast in Fig. 8.2-c which shows bright

linear features inclined with respect to the sample surface. The metallic linear features in

the IOZ explain that EDS detects Cu (and Ni) in surface regions in-between the Cu-particles

at the surface (see above). The increase in sample thickness by development of the IOZ has

led to a crack at the corners of the sample (see Fig. 8.2-b/c). As demonstrated by EDS

elemental mapping (Fig. 8.2-c) Cu and Ni are present, or enriched (Cu), inside these cracks,

while the oxide-forming elements Zr, Al, and Ti are absent.

Fig. 8.2. (a) Secondary electron (SE) micrograph of the top surface morphology of oxidized

Zr51.3Al8.5Cu31.3Ni4Ti4.9. (b) Cross-sectional light optical microscopy of the surface adjacent oxidation zone

obtained from two areas showing metallic linear features inclined (marked by red arrows) with respect to the

sample surface. (c) Cross-sectional BSE micrographs and EDS elemental mapping acquired from the corner

of the oxidized sample. Non-oxidized copper at the surface and inside the crack are marked by white arrows.

Manuscript IV

133

Evaluation of X-ray lattice strains with the sin2ψ method

The lattice strain, ,

hkl

, as experienced by the family of lattice planes {hkl} in a direction

defined by the rotation angle, , and the tilt angle, , and caused by a biaxial state of stress,

( , 3)ij i j (plane stress), in the surface region of the thermochemically oxidized BMG

obeys [26] (Fig. 8.3):

, 0 2 2 21, 2 11 12 22 1 11 222

0

cos sin 2 sin sin

hkl hkl

hkl hkl hkl

hkl

d dS S

d

8-1

where ,

hkld and 0

hkld are the strained and strain-free interplanar lattice spacing,

respectively. 1

hklS and 122

hklS are the X-ray elastic constants (XECs).

Fig. 8.3. Definition and notation of lattice strain measurement directions, i.e. and ψ with respect to the

system of principal stresses in the thermochemically oxidized bulk metallic glass surface with normal vector

𝑛.

The interplanar lattice spacings (111

,d ) as determined from the X-ray experiments are

presented vs. sin2ψ for two mutually perpendicular directions ( 0 and 90 ) in Fig.

8.4. Although the data shows a trend towards slight curvature3, they obey a straight line

3 Several origins can be given for curvature. The most obvious are i) a change in the

magnitude of the biaxial stress with depth, which for the observed curvature would imply


134

(R2>0.99 for 0 and R2>0.98 for 90 ) within experimental accuracy (for

2 0.03 and 2 30 this is 0.001dd

). The slopes in Fig. 8.4 are identical,

which implies that 11 22 . Furthermore, it was verified from measurements for =180°

and =270° that the slight curvature is not a consequence of the presence of a shear stress

12 . Since 11 22 and 12 0 , Eq. 8-1 can be simplified to [26]:

212 / / 1 / /2

sin 2hkl hkl hklS S 8-2

where / / is the in-plane rotationally symmetric stress. From Eq. 8-2, it follows that the so-

called strain-free direction is2 1

0 122

2sin

hkl

hkl

S

S . Then, a value for strain-free lattice

parameter, 0

hkld , is obtained for interpolation in Fig. 8.4 for 2

0sin .

Fig. 8.4. Lattice spacing 111

,d (Å) of the main diffraction peak of c-ZrO2 as a function of sin2ψ for two

rotation angles, 0° and 90°, respectively. The corresponding strain-free direction sin2ψε=0 is specified

in the plots.

that the biaxial stress becomes more compressive with depth, and ii) the presence of a

component 33 deeper in the material or iii) the presence of a shear stress 12 (see also

Discussion).

Manuscript IV

135

Considering that, at least initially, the IOZ consists of crystalline ZrO2 crystals dispersed in

an amorphous matrix rather than a polycrystalline aggregate of only ZrO2, the XECs were

obtained from the single-crystal elastic constants (SECs) for ZrO2, assuming elastic

interaction according to Voigt, implying that all crystallites experience the same strain.

Table 8.2 provides the SECs and, thus calculated, XECs for cubic ZrO2 [39] together with

the macroscopic Poisson ratio (ν) and Young’s modulus (E).

Table 8.2. The single crystal elastic coefficients, Cij [40], calculated X-ray elastic constants (XEC) based on

Voigt assumption for a cubic lattice structure, Poisson ratio, and Young’s modulus of c-Zirconia [41].

C11 (GPa) C12 C44 S1 (10-6 MPa-1) (1/2)S2 (10-6 MPa-1) ν E (GPa) c-ZrO2 401 96 56 -1.20 5.29 0.32 233

Using the XECs from Table 1 the following stress values were calculated from the slopes

in Fig. 8.4: -1,543±63 MPa (=0°) and -1,564±91 MPa ( =90°). These stress values are

equal within experimental accuracy and can be converted into a macroscopic strain, / / ,

(rather than a lattice strain) from:

/ / / /

1

E

8-3

Using Young’s modulus and Poisson constant in Table 8.2, gives 3

/ /( 4.54 0.23) 10

.

Surface strain relaxation by annular FIB milling

Fig. 8.5 shows a series of SEM micrographs acquired from the annular incremental FIB

milling. The dependence of the strain relaxation profile on the ring-core geometry was

studied for cylinder diameters of 4, 6, and 8 μm. Different diameters were chosen to verify

the reproducibility of the stress relaxation for different diameters.


136

Fig. 8.5. SEM micrographs acquired from the incremental ring-core FIB milling process: (a) a 8×8 μm2 Pt-

speckle pattern (reference pattern), (b) intermediate milling step after 92 ion-milling cycles, (c) final milling

step after 184 milling cycles and (d) an overview of the area where the FIB milling process was performed at

the surface region of the oxidized BMG. The final FIB milling depth in this case is ~ 9.5 μm.

Comparing the SEM images before and after each milling cycle in the DIC MATLAB

scrip, DIC post-processing was carried out to convert the image pixel displacement into in-

plane full-field strain. The main DIC analysis was done on the basis of the fine grid of

facets4 shown in Fig. 8.6-a. Each square facet (as for instance a yellow square in Fig. 8.6-

a) contained 5-6 Pt-speckles and the tracked DIC markers (green “+” symbols in Fig. 8.6-

a) were defined with respect to the speckles within the squares. The pixel displacements of

the DIC markers vs. their pixel positions in the X- and Y-directions defined in Fig. 8.6-a,

are given as examples in Fig. 8.6 (b-c). For quantification of the surface-strain relief, a

linear least-squares fit was taken through the cloud of displacements (Figs. 8.6-b,c); the

slope of the linear regression corresponds to the one-dimensional surface strain relief in X-

4 Each green DIC marker shown in Fig. 8.6-a corresponds to the center of each square facet

defined in the region of interest at the surface of the pillar. In addition, for the sake of better

visualization, a yellow square facet which contains multiple pixel points is presented.

Manuscript IV

137

or Y-direction for a specific milling depth. The standard error in the slope of the linear fit

was taken as positive and negative error estimates for the strain values plotted in Fig. 8.6-

d for the sequence of milling steps. For a more accurate assessment of the residual strain,

poorly tracked DIC markers (or outliers) were removed, as suggested in Ref. [30]. A typical

measured surface strain relaxation profile as a function of number of FIB milling cycles

(SEM image No.), i.e. the accumulated milling depth, is presented in Fig. 8.6-d.

Evidently, the strain relaxations in X- and Y-directions converge to the same value after

about 20 milling cycles. Differences in the first 20 cycles, showing an initial negative strain

relief in the Y-direction, i.e. shrinkage rather than expansion, are attributed to edge effects

associated with focused Ga+ milling (see discussion).

The evolution of the mean engineering surface strain relaxation with normalized FIB

milling depth (h/D) for 4, 6, and 8 μm diameter ring-core geometries are collected in Fig.

8.7. The turquoise error bands reflect the standard deviation relative to the arithmetic

average of five FIB-DIC measurements for each of the 3 pillar sizes (altogether 15 pillars

were measured). The character of the in-plane strain follows from the sign of the strain

relief curve: positive values correspond to the relaxation of compressive residual stresses,

while negative values correspond to the release of tensile residual stresses. Evidently, the

strain relief profiles consistently reveal the relaxation of compressive strains and attain a

constant level, indicating that complete strain relief around the pillar is reached. This would

be expected for a milling depth, h, that is at least equal to the pillar diameter, D, (h≥D). It

is noted that a marginal drop in strain value in the beginning of the incremental FIB-milling

process is observed, irrespective of the pillar size. This initial drop is the more pronounced,

the smaller the pillar diameter.


138

Fig. 8.6. Steps required for DIC analysis to compute the full-field strain relief in the incremental ring-core

FIB milling methodology. (a) DIC markers (green “+” symbols) with respect to the pattern of Pt-speckles.

The plots of DIC marker displacements in two mutually perpendicular directions (b) X (red) and (c) Y (blue)

versus the pixel position (with respect to the initial position in the SEM image). The slope of a linear fit

through the cloud of DIC markers provides the surface strain relief estimated for the reached FIB milling

depth. These two plots were obtained after 60 FIB milling cycles for a 4 µm pillar. (d) Measured surface

strain relief profile against number of FIB milling cycles (SEM image No.) for the in-plane X- and Y-

directions. The black curve represents the average of the relaxation profiles in X- and Y-directions. The error

bars represent the standard error in the slopes through the clouds of DIC markers (see b and c).

The in-plane macroscopic strains derived from the incremental ring-core method FIB-DIC

tests at different surface locations, representing the saturation levels in Figs. 8.7, are

collected in Table 8.3.

Manuscript IV

139

Table 8.3. Average in-plane strains, / / , in the IOZ as determined with incremental ring-core milling using

different pillar diameters, D. For each pillar diameter value the strain value represents the average over X-

and Y-directions for 5 pillars.

D (µm) / /

4 37.0 1.5 10

6 34.7 1.0 10

8 35.3 0.6 10

The strain values presented in Table 8.3 are consistent within experimental accuracy. As

reflected by the standard deviations in Table 8.3, the repeatability of the determined strain

value becomes better for larger pillar diameter. Excellent agreement is obtained between

the strains for the 6 and 8 µm pillars and the macro-strains evaluated from the lattice strains

(cf. section 8.3.2).

Fig. 8.7. Evolution of surface strain relaxation with normalized FIB milling depth (h/D) obtained from FIB-

DIC measurement using ring-core geometry of 4, 6, and 8 micron pillars. The black data points indicate the

average strain values obtained from five FIB-DIC measurements performed at different oxidized surface sites

in X- and Y-directions (cf. Fig. 8.6-a). The error band represents the associated standard deviation calculated

based on five FIB-DIC measurements for each ring-core milling geometry.

Discussion

140

8.4 Discussion

Microstructure in the internal oxidation zone

Long-term 60 h exposure of Zr51.3Al8.5Cu31.3Ni4Ti4.9 BMG at a low temperature of 600 K

(<Tg) to an extremely low chemical oxygen potential (pO2≈ 10-41 bar) leads to internal

oxidation under the development of crystalline c-ZrO2. This is consistent with the huge

affinity of Zr to oxygen, as reflected by an equilibrium oxygen partial pressure as low as

pO2≈10-85 bar for the formation of ZrO2 from Zr at 600 K. In the IOZ metallic lines inclined

with respect to the specimen surface develop, while crystalline Cu-particles form at the

sample surface and inside corner cracks. No indications for oxidized copper were found,

which is consistent with an equilibrium pO2≈10-22 bar at 600 K for the formation of Cu2O

from Cu (i.e. the Cu-oxide with most negative Gibbs energy of formation). The

development of an IOZ is consistent with those described in Refs. [24–25]. For the present

case, however, the development of an outer oxidation zone was prevented by reducing pO2

to a pressure below that necessary to oxidize Cu. Hence, finely distributed non-oxidized

Cu particles are observed at the surface of the oxidized BMG and crack surfaces. The

driving force for this outward flux of Cu is provided by the compressive stress developing

during formation of the IOZ.

The internal oxidation zone develops under the influence of inward diffusion of oxygen.

The thickness of the IOZ is determined by diffusion of oxygen through the IOZ and the

provision of oxygen at the surface. Recognizing the extremely low pO2, the

rate-determining step in oxidation is most likely the supply of atomic oxygen to the surface

rather than solid state diffusion of oxygen species (cf. [24]). The presence of metallic linear

features inclined with respect to the surface normal of the sample is interpreted as decorated

shear bands. Shear banding is considered to develop in the BMG underneath the IOZ as a

consequence of stresses caused by volume expansion associated with ZrO2 formation in the

BMG. The presence of shear bands is suggested to provide locations for the segregation

(and possibly crystallization) of non-oxidizing (noble) species, i.e. Cu (and Ni), while the

oxide forming elements Zr, Al and Ti react with oxygen on passing of the penetrating IOZ.

It is noted that the excess free volume in the shear bands provides a mechanism for fast

diffusion of Cu (and Ni) [42]. Since also Al and Ti have a huge affinity for oxygen, it is

assumed that these species oxidize along with ZrO2. Ti and Al have been reported to be

Manuscript IV

141

dissolvable in c-ZrO2 (cf. [43–44]). Recognizing that the contents of Al and Ti are much

lower than the Zr content and that no X-ray diffraction evidence was obtained for the

presence of Al- or Ti-oxides, it is likely that Ti and Al are indeed present in ZrO2. Thus,

the components that have a huge affinity for oxygen are assumed to be present as (Zr, Al,

Ti)O2, implying that 64.7 % of the metal atoms in the IOZ is bound as oxide.

Strain and stress in the internal oxidation zone

The presence of compressive residual stresses within the IOZ was demonstrated ex-situ

(post-mortem) after the oxidation experiment and after cooling to room temperature, with

two independent techniques. The X-ray diffraction sin2ψ method revealed the presence of

lattice strains in the c-ZrO2 particles. These lattice strains are invoked by a state of

rotationally symmetric biaxial compressive residual stresses parallel to the surface in the

particles. The incremental ring-core FIB milling method demonstrated that the evolving

surface strains with increasing ring depth are a consequence of the relaxation of rotationally

symmetric compressive volume strains in the IOZ. The absolute values of the macroscopic

strains evaluated from these independent techniques are consistent with each other for the

larger pillar diameters 6 and 8µm. The two independent methods are further discussed

separately below.

Stress and strain determined with X-ray diffraction sin2ψ method

X-ray diffraction analysis provides diffracted-intensity weighted information, which

implies that the lattice spacings determined in the sin2ψ method represent average values

over a certain depth range. The depth range depends on the diffraction geometry and for

the current experiments the so-called information depth, , i.e. the depth range from which

63% of the diffracted intensity originates, follows from [26]:

sin cos

2

8-4

where µ is the linear absorption coefficient of the applied X-radiation in the analyzed

material (here the IOZ). For the present experiments, this implies that for the 111 peak of

c-ZrO2 the information depth ranges from 4.5 µm (half the thickness of the IOZ) for =0

to 2.3 µm for =60°. Considering the variation in information depth, the slight curvature

of the sin2ψ plots in Fig. 8.4 could be explained as follows. Assuming that the curvature is

due to a gradient in the compressive biaxial residual stress, the compressive stress would

Discussion

142

be larger (more negative) deeper in the IOZ. Assuming a gradient in the stress component

perpendicular to the surface, such that it vanishes at the surface (no stress can exist

perpendicular to a free surface), would correspond to a tensile component parallel to the

surface normal deeper in the IOZ. Alternatively, a gradient in the oxygen content in ZrO2,

such that ZrO2 is more oxygen-rich closer to the surface. The stress-depth distribution was

not quantified for the current data, because the curvature is not significant within

experimental accuracy.

The values obtained for the lattice-plane specific strains in c-ZrO2 were related to the

residual stress in this phase under the assumption that all particles experience the same

strain, as would be the case if they are embedded in a homogenous matrix. From the thus

obtained stress value / / ( 1,554 77) MPa, the macroscopic strain in c-ZrO2 was

obtained as3

// ( 4.54 0.23) 10 .

Stress and strain determined by incremental ring-core milling

The incremental ring-core milling technique provides strain redistribution data for the

volume in the free-standing cylinder that is created. This could be considered as a column

that is clamped at the bottom by a rotationally symmetric biaxial compressive strain.

Investigating different pillar diameters, showed some features that need to be discussed

here. The evolution of the total surface strain relaxation with depth of the ring-core is

presented at the normalized height of the pillar (h/D) in Fig. 8.8. Irrespective of the pillar

diameter, initially a strain relaxation indicates the presence of a tensile stress (see Fig. 8.6-

d for the Y-direction and Fig. 8.7). Similar effects were observed in Ref. [33], but remained

undiscussed. The observed trend that the initial negative strain relief decreases with the

pillar diameter suggests that it is an artefact associated with the pillar surface. It may be a

consequence of ion-beam damage and associated volume effects. It has been reported that

the magnitude and the region of this negative effect is in the range of 10-100 nm and

depends on the milling geometry, material under investigation, and ion beam energy [30].

The FIB-induced change of the state of stress can be minimized by choosing optimal

process parameters, such as low ion-beam current and the right milling geometry, in order

not to directly affect the actual surface strain measurement [34]. In the present case, the

state of stress is likely to be affected by amorphization of ZrO2. Additionally, it is suggested

Manuscript IV

143

that the ion-beam could induce a local transformation of c-ZrO2 to tetragonal or monoclinic

ZrO2, which could affect the strain distribution.

Clearly, excellent agreement is obtained between the data for the strain relaxation

evolutions for 6 µm and 8 µm, while those data for the 4 µm pillar deviate from this. It has

been demonstrated for Si by a combination of experiment and finite-element modeling that

an over-estimate of the surface strain is obtained below a critical cylinder diameter of 2 µm

as a consequence of artefacts induced by Ga+ implantation and surface amorphization [33,

45]. For a material with inferior milling characteristics as compared to Si, as the present

IOZ, a larger critical cylinder diameter can be anticipated. Accordingly, the strain values

obtained for the 4 µm pillar are considered to be affected. Then, taking the accumulated

compressive strains for the 6 and 8 µm pillars as the relaxed macroscopic strain, an average

strain of 3

// ( 5.0 0.3) 10 is obtained, which is in excellent agreement with macro-

strain obtained with X-ray diffraction. Considering the IOZ as a mixture of c-ZrO2 and a

copper matrix, a Poisson ratio of 0.33 and a Young’s modulus of 190 GPa are obtained,

leading to a residual stress in the IOZ of 1,418 85 MPa, using Eq. 8-3.

Fig. 8.8. Total surface strain relaxation profiles with normalized FIB milling depth for 4, 6, and 8 micron

pillars.

Discussion

144

Stress relaxation in the IOZ during oxidation

From the conversion of Zr, Ti, and Al into t-(Zr, Al, Ti)O2, constituting almost 65% of the

atoms in the BMG, the unconstrained volume expansion in the IOZ is 43%. Assuming that

the volume expansion is equally distributed in 3 mutually perpendicular directions, it is

obtained for the strain measured in a direction parallel to the surface -0.143, which is

roughly thirty times as high as the experimentally determined (elastic) macro-strains; this

would correspond to an elastic stress of -36 GPa. Evidently, only a fraction of the volume

expansion associated with internal oxidation is accommodated elastically; the majority has

been accommodated plastically. The mechanisms by which plastic accommodation of the

volume strains can be achieved are the following. Firstly, the shape of the nano-scale Zr-

based oxides could be elongated along the surface normal of the BMG. In a previous

investigation, no indications for the presence of prolate particles in the IOZ were observed

[25].

Secondly, plastic deformation in the unoxidized BMG during oxidation could occur,

recognizing that a ZrCuAl-based BMG has a yield strength of about 1.5 GPa under

compression [46]. The stress state imposed onto the BMG underneath the penetrating IOZ

is biaxial tensile. Plastic deformation of BMGs is associated with shear banding. The linear

metallic features inclined about 60° with the surface as observed in the IOZ (Fig. 8.2-b/c)

are interpreted as evidence that shear banding did occur indeed. Upon passing of the

penetration IOZ, the shear bands are decorated with “noble” elements, because of the larger

free volume in the shear band.

Thirdly, the potentially high compressive stresses associated with volume expansion can

establish a driving force for the outward diffusion of un-oxidized (“noble”) metal atoms of

Cu (and Ni). Similar outward diffusion of “noble” species was observed on internal

oxidation of Ag-In alloys [47]. The diffusivity of Cu is expected to be higher than for Ni

and diffusion is anticipated to be further assisted by the presence of interfaces between

nano-scale Zr- based oxides and the excess free volume in the shear bands (see above and

[42]). This mechanism has eventually led to the development of facetted, crystalline Cu

hillocks at the surface as well as the metallic lines in the IOZ that are inclined with respect

to the surface (see below). Furthermore, this mechanism has led to segregation and

crystallization of Cu (with dissolved Ni) to corner cracks, leading to self-healing [25] and

Manuscript IV

145

the decoration of shear bands. An estimate of the strain relaxation that can be attained by

removing all “noble” species from the BMG would lead to a volume reduction of about

31%. As compared to the above mentioned 43% volume increase on internal oxidation this

would indeed imply a potentially large stress-relaxation mechanism. Assuming that, along

with internal oxidation, all “noble” elements leave the IOZ, the remaining biaxial macro-

strain is -0.04, which is still about 8 times as large as the experimentally determined macro-

strain within the plane of the surface. An overview of the proposed oxidation mechanism

is summarized in Fig. 8.9.

Fig. 8.9. Schematic illustration of the proposed oxidation mechanism for low-pO2 oxidized

Zr51.3Al8.5Cu31.3Ni4Ti4.9 BMG and the corresponding stress relaxation mechanism.

8.5 Conclusion

Oxidation of a Zr51.3Al8.5Cu31.3Ni4Ti4.9 BMG by means of gaseous oxidizing below the

glass transition temperature at an extremely low pO2 leads to the development of an internal

oxidation zone (IOZ) containing c-ZrO2 and crystalline Cu particles on the surface and in

corner cracks. Furthermore, the IOZ contains metallic linear features inclined with respect

to the surface, which are interpreted as shear bands decorated with non-oxidized metallic

species Cu and Ni. The residual stresses developing in the IOZ were investigated using X-

ray diffraction and an incremental ring-core FIB-DIC method. Excellent correspondence

between the two independent strain-determination techniques was obtained. In the XRD-

investigation, the results demonstrate that compressive residual macro-strains of ~ 0.45%

Conclusion

146

have the developed in the crystalline c-ZrO2, corresponding to a residual stress of -1.5 GPa.

The incremental ring-core FIB-DIC method reveals the presence of compressive residual

macro-strains of 0.50% in the internal oxidation zone, corresponding to a residual stress of

-1.4 GPa. The developing residual stresses are much smaller than those expected if 65% of

the metallic constituents (Zr, Al, and Ti) in the BMG forms c-(Zr, Al, Ti)O2, indicating that

stress relaxation has occurred during internal oxidation. The microstructural features

indicate that stress relaxation has occurred by plastic deformation in the BMG underneath

the IOZ and stress-induced diffusion of unoxidized species out of the IOZ, in particular Cu.

After their formation, the shear bands are decorated with metallic species, while the IOZ

penetrates into the BMG. The stress-induced diffusion of metallic species is evidenced by

the Cu-hillocks at the outer surface and Cu/Ni segregation in corner cracks, leading to self-

healing. Stress-induced diffusion of Cu (Ni) is also held responsible for the decoration of

shear bands.

Acknowledgements

The authors sincerely acknowledge the financial support from Villum Fonden (Grant No.

13253). F. Niessen acknowledges the Danish Council for Independent Research (Grant:

DFF-8027-00009B) for financial support. This research used the FEI NanoLab G3 CX

funded by the Australian Research Council (ARC)—Linkage, Infrastructure, Equipment

and Facilities (LIEF) Grant (LE160100063) located at the UOW Electron Microscopy

Centre. The authors would also thank the Institute of Applied Physics, Jiangxi Academy of

Sciences, Nanchang in China for providing the starting material.

Manuscript IV

147

References

[1] W. Klement, R. H. Willens, and P. Duwez, “Non-crystalline structure in solidified

gold-silicon alloys,” Nat. Sci., vol. 187, no. 4740, pp. 869–870, 1960.

[2] A. L. Greer, Metallic Glasses, Fifth Edit., vol. 1. Elsevier B.V., 2014.


[4] W. L. Johnson, “Bulk amorphous metal-An emerging engineering material,” Jom,

vol. 54, no. 3, pp. 40–43, 2002.



[6] C. J. Byrne and M. Eldrup, “Bulk metallic glasses,” Science (80-. )., vol. 321, no.

5888, pp. 502–503, 2008.


vol. 54, no. 3, pp. 321–326, 2006.



[9] H. F. Li and Y. F. Zheng, “Recent advances in bulk metallic glasses for biomedical

applications,” Acta Biomater., vol. 36, pp. 1–20, 2016.



[11] G. Kumar, A. Desai, and J. Schroers, “Bulk metallic glass: The smaller the better,”

Adv. Mater., vol. 23, no. 4, pp. 461–476, 2011.

[12] M. M. Trexler and N. N. Thadhani, “Mechanical properties of bulk metallic glasses,”

Prog. Mater. Sci., vol. 55, no. 8, pp. 759–839, 2010.

[13] C. A. Schuh, T. C. Hufnagel, and U. Ramamurty, “Mechanical behavior of

amorphous alloys,” Acta Mater., vol. 55, no. 12, pp. 4067–4109, 2007.

[14] A. R. Yavari, J. J. Lewandowski, and J. Eckert, “Mechanical Properties of Bulk

Metallic Glasses,” vol. 32, no. August, pp. 635–638, 2007.

[15] R. O. Ritchie, “The conflicts between strength and toughness,” Nat. Mater., vol. 10,

no. 11, pp. 817–822, 2011.

[16] S. V. Ketov et al., “Rejuvenation of metallic glasses by non-affine thermal strain,”

Nature, vol. 524, no. 7564, pp. 200–203, 2015.

[17] Y. Sun, A. Concustell, and A. L. Greer, “Thermomechanical processing of metallic

glasses: Extending the range of the glassy state,” Nature Reviews Materials. 2016.

[18] A. L. Greer and Y. H. Sun, “Stored energy in metallic glasses due to strains within

the elastic limit,” Philos. Mag., vol. 96, no. 16, pp. 1643–1663, 2016.



Conclusion

148

[20] S. V. Madge, D. V. Louzguine-Luzgin, J. J. Lewandowski, and A. L. Greer,

“Toughness, extrinsic effects and Poisson’s ratio of bulk metallic glasses,” Acta

Mater., vol. 60, no. 12, pp. 4800–4809, 2012.

[21] Y. Zhang, W. H. Wang, and A. L. Greer, “Making metallic glasses plastic by control

of residual stress,” vol. 5, no. November, pp. 857–860, 2006.

[22] J. Fan, A. Chen, J. Wang, J. Shen, and J. Lu, “Improved plasticity and fracture

toughness in metallic glasses via surface crystallization,” Intermetallics, vol. 19, no.

10, pp. 1420–1427, 2011.

[23] L. W. Ren et al., “Enhancement of plasticity in Zr-based bulk metallic glasses

electroplated with copper coatings,” Intermetallics, vol. 57, pp. 121–126, 2015.

[24] S. Haratian, F. B. Grumsen, M. Villa, T. L. Christiansen, and M. A. J. Somers,

“Surface hardening by gaseous oxidizing of (Zr55Cu30Al10Ni5)98Er2 bulk-metallic

glass,” J. Alloys Compd., vol. 800, pp. 456–461, 2019.





Evaluation-Application-Assessment. 1997.

[27] P. J. Withers and H. K. D. H. Bhadeshia, “Residual stress part 1 - Measurement

techniques,” Mater. Sci. Technol., vol. 17, no. 4, pp. 355–365, 2001.

[28] Alexander M. Korsunsky, A Teaching Essay on Residual Stresses and Eigenstrains.

2017.

[29] B. Winiarski, A. Gholinia, J. Tian, Y. Yokoyama, P. K. Liaw, and P. J. Withers,

“Submicron-scale depth profiling of residual stress in amorphous materials by

incremental focused ion beam slotting,” Acta Mater., vol. 60, no. 5, pp. 2337–2349,

2012.

[30] A. J. G. Lunt and A. M. Korsunsky, “A review of micro-scale focused ion beam

milling and digital image correlation analysis for residual stress evaluation and error

estimation,” Surf. Coatings Technol., vol. 283, pp. 373–388, 2015.

[31] A. M. Korsunsky, M. Sebastiani, and E. Bemporad, “Focused ion beam ring drilling

for residual stress evaluation,” Mater. Lett., vol. 63, no. 22, pp. 1961–1963, 2009.

[32] S. Keil, “Experimental Determination of Residual Stresses With the Ring‐Core

Method and an on‐Line Measuring System,” Exp. Tech., vol. 16, no. 5, pp. 17–24,

1992.

[33] A. M. Korsunsky et al., “Nanoscale residual stress depth profiling by Focused Ion

Beam milling and eigenstrain analysis,” Mater. Des., vol. 145, pp. 55–64, 2018.

[34] X. Song et al., “Residual stress measurement in thin films at sub-micron scale using

Focused Ion Beam milling and imaging,” Thin Solid Films, vol. 520, no. 6, pp.

2073–2076, 2012.

[35] A. M. Korsunsky, M. Sebastiani, and E. Bemporad, “Residual stress evaluation at

Manuscript IV

149

the micrometer scale: Analysis of thin coatings by FIB milling and digital image

correlation,” Surf. Coatings Technol., vol. 205, no. 7, pp. 2393–2403, 2010.

[36] F. Niessen and M. J. B. Nancarrow, “Computer-aided manufacturing and focused

ion beam technology enable machining of complex micro- And nano-structures,”

Nanotechnology, vol. 30, no. 43, 2019.

[37] F. Di Gioacchino and W. J. Clegg, “Mapping deformation in small-scale testing,”

Acta Mater., vol. 78, no. June 2018, pp. 103–113, 2014.

[38] C. Eberl, R. Thompson, D. . Gianola, and S. Bundschuch, “Digital Image

Correlation and Tracking,” 2010. [Online]. Available:

https://www.mathworks.com/matlabcentral/fileexchange/12413-digital-image-

correlation-and-tracking.

[39] P. Schoderbock, P. Leibenguth, and M. Tkadletz, “Pattern decomposition for

residual stress analysis: A generalization taking into consideration elastic anisotropy

and extension to higher-symmetry Laue classes,” J. Appl. Crystallogr., vol. 50, pp.

1011–1020, 2017.

[40] E. H. Kisi and C. J. Howard, “Elastic Constants of Tetragonal Zirconia Measured by

a New Powder Diffraction Technique,” J. Am. Ceram. Soc., vol. 81, no. 6, pp. 1682–

1684, 1998.

[41] H. M. Kandil, J. D. Greiner, and J. F. Smith, “Single‐Crystal Elastic Constants of

Yttria‐Stabilized Zirconia in the Range 20° to 700°C,” J. Am. Ceram. Soc., vol. 67,

no. 5, pp. 341–346, 1984.

[42] J. Bokeloh, S. V. Divinski, G. Reglitz, and G. Wilde, “Tracer measurements of

atomic diffusion inside shear bands of a bulk metallic glass,” Phys. Rev. Lett., vol.

107, no. 23, pp. 1–5, 2011.

[43] S. Inamura, H. Miyamoto, Y. Imaida, M. Takagawa, K. Hirota, and O. Yamaguchi,

“Formation and hot isostatic pressing of ZrO2 solid solution in the system ZrO2-

Al2O3,” J. Mater. Sci., vol. 29, no. 18, pp. 4913–4917, 1994.

[44] H. J. Goldschmidt, Interstitial Alloys. 1967.

[45] E. Salvati, T. Sui, A. J. G. Lunt, and A. M. Korsunsky, “The effect of eigenstrain

induced by ion beam damage on the apparent strain relief in FIB-DIC residual stress

evaluation,” Mater. Des., vol. 92, no. December 2015, pp. 649–658, 2016.

[46] Z. F. Zhang, J. Eckert, and L. Schultz, “Difference in compressive and tensile

fracture mechanisms of Zr59Cu20Al10Ni8Ti3 bulk metallic glass,” Acta Mater., vol.

51, no. 4, pp. 1167–1179, 2003.

[47] S. Guruswamy, S. M. Park, J. P. Hirth, and R. A. Rapp, “Internal oxidation of Ag-

in alloys: Stress relief and the influence of imposed strain,” Oxid. Met., vol. 26, no.

1–2, pp. 77–100, 1986.

Conclusion

150

Conclusion

151

9 Conclusion

In this Ph.D. project, the surface reaction of bulk metallic glasses (BMGs) exposed to

oxidizing atmospheres imposing different oxygen partial pressures (pO2) at temperatures

below the glass transition temperature (Tg) is studied using in-situ and ex-situ materials

characterization techniques. This is performed to investigate the possibility of surface

engineering of BMGs by modifying their surface-composition by incorporating oxygen.

Moreover, the role of (compressive) residual stresses developed as a consequence of

oxygen dissolution in the surface region of BMGs on the underlying oxidation mechanism

is investigated using in-situ (and ex-situ) X-ray diffraction (XRD) and incremental ring-

core focused ion beam milling together with digital image correlation analysis (FIB-DIC).

For these purposes, three ZrCuAl-based BMGs are used, which are characterized by the

presence of specific elements in their chemistry, i.e. Ag-, Er-, and Ti-containing BMGs.

The research activities provide a large amount of information; the key research findings are

summarized in this chapter. The conclusions are collected and categorized based on the

different classes of analysis, investigations, and experimental works. To find all detailed

conclusions achieved from the experimental works, the reader is referred to individual

result chapters (5-8).

(i) Oxidation-induced surface microstructure in ZrCuAl-based BMGs

The in-situ XRD investigations show that the dissolution of oxygen induces the

development of nano-crystalline ZrO2 at an early stage of oxidation of Ag- and Er-

containing ZrCuAl-based BMGs in the inner oxidation zone (IOZ) at a temperature

about ~ 450-470 K.

Long-term sub-Tg oxidizing at an extremely low pO2 (<10-22 bar) results in the

development of an IOZ while oxidizing at a higher pO2 (>10-22 bar) leads to the

formation of two oxidation zones: and internal and an outer oxidation zone (IOZ and

OOZ, respectively).

The growth of the IOZ is suggested to be rate-controlled by the surface reaction, i.e.

oxygen transfer from the gas to the solid at extremely low pO2 (<10-22 bar). For a high-

pO2 the growth of the (internal) oxidation zone appears rate-controlled by diffusion,

most likely of oxygen through the IOZ.

152

Post (microscopical) studies of in-situ air oxidized Er- and Ag- containing ZrCuAl-

based BMGs demonstrate that IOZ is mainly comprised of tetragonal ZrO2 indicating

Cu enrichments along the (perpendicular or inclined) micro-cracks with respect to the

surface and the OOZ consists mostly of (porous) Cu-based oxides. In the case of Ag-

containing BMG, the OOZ comprises of a stratified distribution Cu oxides, metallic

Ag, and ZrO2.

In the case of Ti-containing BMG, low-pO2 (≈10-41 bar) oxidizing treatment at 600 K

for 60 h leads to the development of an internal oxidation zone comprising of cubic

ZrO2, metallic inclined linear features and expelled individual Cu particles at the

surface.

The presence of Cu-based oxides and metallic “noble” elements like Cu (and Ag) in the

OOZ are attributed to the outward diffusion of these metallic components from the

developing IOZ as a consequence of compressive stress build-up caused by the

formation of ZrO2.

Applying oxidizing at temperatures close to the Tg is found to be accompanied by partial

devitrification of the (Er-containing ZrCuAl-based) BMG substrate. It is found that the

IOZ grows as a result of internal oxidation and partial crystallization of the amorphous

substrate.

(ii) Self-healing of ZrCuAl-based BMGs during exposure to oxidation

The newly created surfaces developed by cracking/shearing in the IOZ (and BMG), are

effectively repaired and pacified by the crystallization and segregation of the noble

elements (mostly Cu), thus showing a surprising self-healing effect.

The mechanism behind the self-healing phenomenon is suggested to be compressive

stress-assisted diffusion of the noble elements to the free surfaces, where they segregate

as crystalline metals and thereby seal the cracks.

It is observed that in the case of in-situ air-oxidized Er-containing BMG, oxidation of

Cu within the (self-healed) micro-cracks could occur if the local pO2 in these areas is

higher than ~10-22 bar. It is inferred this effect can cause Cu-oxide wedge formation

inside the cracks leading to the crack growth as well as the branching of the cracks.

(iii) The effect of oxygen dissolution on surface hardness of ZrCuAl-based BMGs

The surface hardness of ZrCuAl-based BMG can be improved by the oxidizing

treatment in controlled atmospheres below the Tg, where the surface-hardened case

depth is highly dependent on the applied pO2.

Conclusion

153

The hardness in the oxygen-containing case increased significantly from approximately

~ 600 HV in the core of the un-oxidized Er-containing ZrCuAl-based BMG to ~ 1200

HV.

(iv) Strain, stress, and stress relaxation in low-pO2 oxidized Ti-containing BMG

Surface engineering by thermochemical gaseous oxidizing is demonstrated to be an

effective method for introducing compressive residual stresses in the surface region of

the ZrCuAl-based BMGs.

The determination of residual stresses (and strain) developed in the hardened-low pO2

oxidized Ti-containing ZrCuAl-based BMG using conventional XRD sin2ψ and

incremental ring-core FIB-DIC methods show an excellent correspondence in the

measured values between these two independent techniques.

The residual macro-strain sin2ψ analysis conducted on the (111) reflection of crystalline

c-ZrO2 reveals the existence of compressive macro-strains of ~ 0.45%, corresponding

to residual stress of -1.5 GPa. The surface strain relaxation measured by the FIB-DIC

method shows the presence of compressive residual macro-strains of 0.50% in the

internal oxidation zone, corresponding to residual stress of -1.4 GPa.

The high compressive residual stresses expected to develop as a consequence of volume

expansion associated with the formation of Zr(Al and Ti)O2 in the IOZ are noticeably

relaxed by the plastic accommodation of the volume strains. The occurrence of stress

relaxation during the evolution of the IOZ is accomplished by: (1) plastic deformation

in the underlying BMG as is evidenced by the formation of shear bands and (2) stress-

induced outward transport of noble elements from IOZ leading to the development of

Cu-hillocks at the outmost surface and decoration of shear bands/corner cracks with un-

oxidized Cu.

(v) In-situ residual stress evolution and stress relaxation in Er-containing BMG

during air-oxidation

The conversion of Zr to ZrO2 at an early stage of oxidation leads to the build-up of

compressive residual stresses of ~ -1.4 GPa in the surface region.

The determination of the state of stress during in-situ air-oxidation of the BMG shows

that compressive residual stresses induced by the formation of t-ZrO2 become

significantly lower, dropping from ~ -1400 MPa to ~ -400 MPa on continued heating

to 600 K. It is anticipated that the stress relief occurs because of: (1) Cu (and Ni)

redistribution and outward diffusion of these noble elements (with less affinity for

154

oxygen than Zr) from developing IOZ to the free surfaces and in particular (2)

perpendicular crack formation for compensating tensile stresses in the BMG.

The in-situ XRD lattice strain analysis shows that negligible compressive residual

stresses as compared to the formation ZrO2 are developed by Cu-based oxides

formation, i.e., m-CuO and c-Cu2O due to their volume increase. The average growth

compressive residual stresses for the reflections (111) CuO and (111) Cu2O are ~100

MPa and ~50 MPa, respectively.

Further work

155

10 Further work

The present Ph.D. work investigates the surface response of the multicomponent ZrCuAl-

based BMGs to the various oxygen-containing atmosphere at elevated temperatures. The

emphasis is on surface microstructure evolution, surface hardening, introducing

compressive residual stresses in the surface region of the BMGs, and more specifically

determining the mechanisms responsible for the oxidation of such defect-free materials. In

addition to the current understanding of how oxygen dissolution in the ZrCuAl-based

glassy-structured materials influences their near-surface zone microstructure, some further

investigations are suggested to be applied in continuation of the presented experimental

results where some of them are listed as follows:

In the present work, it is suggested that surface engineering by low temperature gaseous

thermochemical oxidizing can be used as an effective method to introduce compressive

residual stresses in the surface region of ZrCuAl-based BMGs, aiming at suppressing

shear band localization during mechanical loading, i.e. delaying crack initiation.

Nevertheless, despite a significant improvement of the surface hardness by gaseous

oxidizing, at the same time, thermal oxidizing may have a detrimental effect on the

fracture toughness (and plasticity) of a BMG substrate due to annihilation of the free

volume in the glassy alloy, that was stored during the transition from super-cooled

liquid to glass. It is known that the annealing of BMGs induces structural relaxation,

reaching the states of a lower degree of structural disorder, which ultimately results in

drastic brittleness. As opposed to structural relaxation, rejuvenation of BMGs is

possible by the thermal cycling of BMGs at cryogenic temperatures. The deliberate

structural changes induced by cryogenic rejuvenation increase the stored energy

(enthalpy of relaxation) as a consequence of inhomogeneous thermal strain distribution

in the glass. Therefore, it is of interest to systematically investigate the effect of cryo-

thermal cycling on a thermochemically surface-engineered BMG to reverse the

annealing-induced brittleness.

The microscopical characterization combined with spectroscopy of the investigated

oxidized ZrCuAl-based BMGs demonstrates the presence of Al (and Ti) in the inner

oxidation zone (IOZ). However, no indications of the Al- (and Ti-) based oxides in the

IOZ are detected applying X-ray (and electron) diffraction techniques. Hence,

supplementary works on investigating the distribution of these elemental constituents

156

with a high affinity for oxygen using atom probe tomography (APT) would be

beneficial to make a clear interpretation of this matter. Moreover, it is found that long-

term oxidizing of Ti-containing ZrCuAl-based BMG exposed to an extremely low pO2

leads to the formation and stabilization of cubic ZrO2 at low temperature below Tg,

while in the case of sub-Tg oxidized Er-containing BMG, tetragonal ZrO2 is developed

in its surface region. It is anticipated that, in the case of low-pO2 oxidized Ti-containing

BMG, the presence of Al and Ti in the chemical composition of the BMG could

promote the stabilization of c-ZrO2 at low temperatures. Therefore, a fundamental study

should be carried out in order to understand how BMG chemical composition can

influence the lattice structure of ZrO2.

In the present work, it is shown that the build-up of compressive residual stresses as a

consequence of volume expansion associated with ZrO2 (Al and Ti oxides) formation

in low-pO2 oxidized Ti-containing BMG, leads to the initiation of shear bands inclined

with respect to the surface. On the other hand, microscopical investigations on an in-

situ air-oxidized Er-containing BMG reveals the development of (micro-) cracks

perpendicular to the surface of the BMG. It is assumed that the micro-cracks (shear

bands and perpendicular cracks) initiation occurs in an early stage of oxidation in the

BMG before the internal oxidation zone (IOZ) grows. However, it is still not well-

understood that which parameters affect the type of crack patterns. It is anticipated that

this can either be as a consequence of intrinsic properties of the investigated oxidized

BMGs (for instance difference in their fracture toughness) or the type of ZrO2 (cubic or

tetragonal lattice structure). Therefore, in order to elucidate these unaddressed points,

further experimental works are required to be conducted.

In this project, it is hypothesized that due to the presence of highly-oxidizable elemental

constituents in the chemical composition of Zr-based BMGs like Zr and Al, gaseous

oxidizing could be used as an efficient method for surface engineering. It addition to

the incorporation of oxygen, there is a largely unexplored potential to study the

possibility of dissolving nitrogen and hydrogen in the surface of the BMGs at

temperatures below their glass transition temperature (Tg).

In the case of this study, it is attempted to apply oxidizing at a temperature preferably

below Tg to avoid devitrification of the BMG substrate. Nevertheless, thorough

investigations require to be performed to understand the effect of various temperatures

at different oxygen potential on surface-oxidation microstructure as well as the

Further work

157

formation of micro-cracks (and shear bands). Moreover, in-depth understanding of the

effect of oxidizing parameters, including external parameters (time, temperature, and

oxygen partial pressure) and internal parameters (chemical composition of the BMG),

on the development of micro-cracks (and shear bands) during oxidation of the multi-

component ZrCuAl-based BMGs is not fully achieved based on the current results of

the Ph.D. project. For this, a more detailed stepwise oxidation at different temperatures

and oxygen potentials combined with microstructural observation after each step

(during heating, isothermal holding, cooling to ambient temperature) is strongly

suggested.

The focus of this Ph.D. work is mainly on the oxidation behavior and the role of residual

stresses developed during oxidation of ZrCuAl-based BMGs on the observed surface

oxidation zone(s) microstructure using both in- and ex-situ techniques. This study can

also be extended and applied to the other BMGs like Ti-, Fe-, Cu-based BMGs, etc. to

shed light on the oxidation governing mechanism in these glassy alloying systems.

DTU Mechanical EngineeringSection of Materials and Surface EngineeringTechnical University of Denmark

Produktionstorvet, Bld. 425DK-2800 Kgs. Lyngby Denmark

Tlf.: +45 4525 2205Fax: +45 4525 1961

www.mek.dtu.dk

May 2020

ISBN: 978-87-7475-598-2

surface engineering of zrcual-based bulk metallic glasses ......e-mail: [email protected] technical...

Documents